Online shopping is getting a new AI twist. OpenAI has introduced virtual try-on in ChatGPT, allowing users to upload a selfie or full-body photo and see how selected clothes and accessories could look on them before making a purchase.&nbsp;The feature is rolling out globally alongside a new Favorites tool, taking ChatGPT beyond product recommendations and towards a more visual shopping experience.<h2>How ChatGPT virtual try-on works</h2>Users can upload a photo of themselves and select the new Try On option when it appears with eligible shopping results. ChatGPT then generates an image showing the selected item on the user. There is another way to use the feature. Users can upload a screenshot of a product from a shopping website and ask ChatGPT to virtually try it on them. The feature is powered by OpenAI's ChatGPT Images 2.5 model. OpenAI says the model is better at handling lighting and textures, follows image-editing instructions more accurately and generates images with lower latency. This could be particularly useful for people who are unsure whether a particular style, colour or item will suit them.&nbsp;<h2>From one product to an entire outfit</h2>The tool is not limited to trying on a single item. Users can describe the kind of look they want and ask ChatGPT to find clothes and accessories that fit that style. They can also upload an image of a celebrity outfit and ask ChatGPT to find similar products that are available to buy. That turns ChatGPT into more than a traditional shopping search tool. Instead of simply finding products, it can help users explore how different pieces could work together visually. However, shoppers should remember that an AI-generated preview is not a guarantee of how an item will actually fit or look in real life. Factors such as sizing, fabric, cut and lighting can affect the final appearance.&nbsp;<h2>ChatGPT Favorites lets shoppers save products</h2>OpenAI is also introducing Favorites, allowing users to save products they discover while shopping through ChatGPT. Saved products will appear in the Library, alongside images generated through virtual try-on. Users can therefore return to products they liked without having to search for them again. Together, the two features show how OpenAI is expanding ChatGPT's role in shopping from answering questions and finding products to helping users visualise, compare and save potential purchases.

OpenAI has introduced virtual try-on and Favorites in ChatGPT. Users can upload selfies or full-body photos to see how eligible clothes and accessories may look on them, while Favorites lets shoppers save products for later.

Online shopping is getting a new AI twist. OpenAI has introduced virtual try-on in ChatGPT, allowing users to upload a selfie or full-body photo and see how selected clothes and accessories could look on them before making a purchase. The feature is rolling out globally alongside a new Favorites tool, taking ChatGPT beyond product recommendations and towards a more visual shopping experience. How ChatGPT virtual try-on works Users can upload a photo of themselves and select the new Try On option when it appears with eligible shopping results. ChatGPT then generates an image showing the selected item on the user. There is another way to use the feature. Users can upload a screenshot of a product from a shopping website and ask ChatGPT to virtually try it on them. The feature is powered by OpenAI's ChatGPT Images 2.5 model. OpenAI says the model is better at handling lighting and textures, follows image-editing instructions more accurately and generates images with lower latency. This could be particularly useful for people who are unsure whether a particular style, colour or item will suit them. From one product to an entire outfit The tool is not limited to trying on a single item. Users can describe the kind of look they want and ask ChatGPT to find clothes and accessories that fit that style. They can also upload an image of a celebrity outfit and ask ChatGPT to find similar products that are available to buy. That turns ChatGPT into more than a traditional shopping search tool. Instead of simply finding products, it can help users explore how different pieces could work together visually. However, shoppers should remember that an AI-generated preview is not a guarantee of how an item will actually fit or look in real life. Factors such as sizing, fabric, cut and lighting can affect the final appearance. ChatGPT Favorites lets shoppers save products OpenAI is also introducing Favorites, allowing users to save products they discover while shopping through ChatGPT. Saved products will appear in the Library, alongside images generated through virtual try-on. Users can therefore return to products they liked without having to search for them again. Together, the two features show how OpenAI is expanding ChatGPT's role in shopping from answering questions and finding products to helping users visualise, compare and save potential purchases.

OpenAI has introduced virtual try-on and Favorites in ChatGPT. Users can upload selfies or full-body photos to see how eligible clothes and accessories may look on them, while Favorites lets shoppers save products for later.

ChatGPT gets virtual try-on feature: You can now see how clothes look on you before buying

A senior OpenAI safety employee has resigned after three-and-a-half years at the company, warning that its “culture is broken” and that the risks from increasingly capable AI systems are becoming harder to manage.&nbsp;David Robinson said he worked on safety reports accompanying major OpenAI product launches. In an essay published by The Atlantic, he argued that the industry needs to rethink how it develops frontier AI, rather than relying mainly on rules added after problems emerge. Robinson's departure comes as OpenAI faces growing scrutiny over the behaviour of AI agents and the safeguards designed to control them.&nbsp;<h2>Why Robinson says OpenAI needs a different approach</h2><h2>&nbsp;</h2>Robinson said OpenAI has relied heavily on what it calls “iterative deployment” releasing systems, finding problems and then improving their safeguards. He argues that approach becomes increasingly risky as AI models grow more capable. According to Robinson, failures are inevitable under such a system, but the consequences could become much more serious as AI agents gain the ability to carry out complex tasks with less human involvement. He pointed to the reported breach of Hugging Face systems by OpenAI agents and other discoveries involving rogue agent behaviour as examples of why stronger protections are needed. Robinson believes frontier AI companies should instead operate more like nuclear power plants or major airports, where multiple layers of safeguards are designed to prevent a single human mistake from becoming a disaster.&nbsp;<h2>OpenAI responds to the criticism</h2>OpenAI disputed the suggestion that it is ignoring safety. Company spokesperson Drew Pusateri said OpenAI is strengthening security in its research and testing environments and is working to ensure that increasingly capable models remain manageable and secure. The company also said it is expanding third-party evaluations and improving real-time monitoring so potentially concerning behaviour can be identified earlier. Robinson, however, believes technical safeguards alone are not enough.&nbsp;<h2>Why AI alignment is becoming a bigger issue</h2>Robinson also called for greater attention to AI alignment the challenge of ensuring advanced AI systems behave in ways that remain consistent with human goals and values. He argued that current methods for measuring whether AI systems match human values remain limited. His concerns echo warnings from other former AI researchers, including Jacob Coxon, who has also criticised the industry's approach to safety. Robinson said he ultimately concluded that external incentives for safety may be necessary because employees working inside frontier AI companies are often too focused on rapidly developing and deploying new systems to make fundamental cultural changes. His resignation adds another prominent voice to a growing debate over whether the AI industry can safely move fast enough as its technology becomes more powerful.

OpenAI safety veteran David Robinson has resigned after 3.5 years, saying the company’s culture is “broken”. He warns that increasingly capable AI systems require stronger safeguards and argues that frontier AI companies need deeper cultural and external safety incentives.

A senior OpenAI safety employee has resigned after three-and-a-half years at the company, warning that its “culture is broken” and that the risks from increasingly capable AI systems are becoming harder to manage. David Robinson said he worked on safety reports accompanying major OpenAI product launches. In an essay published by The Atlantic, he argued that the industry needs to rethink how it develops frontier AI, rather than relying mainly on rules added after problems emerge. Robinson's departure comes as OpenAI faces growing scrutiny over the behaviour of AI agents and the safeguards designed to control them. Why Robinson says OpenAI needs a different approach Robinson said OpenAI has relied heavily on what it calls “iterative deployment” releasing systems, finding problems and then improving their safeguards. He argues that approach becomes increasingly risky as AI models grow more capable. According to Robinson, failures are inevitable under such a system, but the consequences could become much more serious as AI agents gain the ability to carry out complex tasks with less human involvement. He pointed to the reported breach of Hugging Face systems by OpenAI agents and other discoveries involving rogue agent behaviour as examples of why stronger protections are needed. Robinson believes frontier AI companies should instead operate more like nuclear power plants or major airports, where multiple layers of safeguards are designed to prevent a single human mistake from becoming a disaster. OpenAI responds to the criticism OpenAI disputed the suggestion that it is ignoring safety. Company spokesperson Drew Pusateri said OpenAI is strengthening security in its research and testing environments and is working to ensure that increasingly capable models remain manageable and secure. The company also said it is expanding third-party evaluations and improving real-time monitoring so potentially concerning behaviour can be identified earlier. Robinson, however, believes technical safeguards alone are not enough. Why AI alignment is becoming a bigger issue Robinson also called for greater attention to AI alignment the challenge of ensuring advanced AI systems behave in ways that remain consistent with human goals and values. He argued that current methods for measuring whether AI systems match human values remain limited. His concerns echo warnings from other former AI researchers, including Jacob Coxon, who has also criticised the industry's approach to safety. Robinson said he ultimately concluded that external incentives for safety may be necessary because employees working inside frontier AI companies are often too focused on rapidly developing and deploying new systems to make fundamental cultural changes. His resignation adds another prominent voice to a growing debate over whether the AI industry can safely move fast enough as its technology becomes more powerful.

OpenAI safety veteran David Robinson has resigned after 3.5 years, saying the company’s culture is “broken”. He warns that increasingly capable AI systems require stronger safeguards and argues that frontier AI companies need deeper cultural and external safety incentives.

OpenAI safety expert quits, calls for nuclear-style safeguards for AI

In September, Australian Prime Minister Anthony Albanese revealed that an autonomous OpenAI AI agent unauthorisedly breached a government portal in June. OpenAI apologised to the Australian government on Tuesday (Sep 29), which involved Medicare, the country’s universal health insurance scheme. The autonomous AI systems accessed the Medicare Statistics Reporting Service portal, administered by Services Australia.&nbsp;ABC News's Cam Wilson shared the email sent by OpenAI to the Australian government's Services Australia to reveal what had happened. The letter reads that the "security vulnerability" was identified "during our review of OpenAI Model activity…" It goes on to say, "An OpenAI model identified a way to make the server carry out instructions sent through the public reporting interface, without a private account or password."&nbsp;The agent was able to "read portions of internal program files and settings, obtain a list of files, and create and read back a small test file on the server." OpenAI said that it reviewed the incident but found "no evidence that the model accessed patient-level records, personal information or credentials; deleted data; or established ongoing access."Also Read: <a href="https://www.wionews.com/world/openai-admits-ai-agent-scaled-the-fence-into-australia-government-system-apologises-we-are-sorry-1790658789423">OpenAI admits AI agent ‘scaled the fence’ into Australia government system, apologises: ‘We are sorry’</a>&nbsp;It then tries to work its way around the issue, by portraying itself as the saviour who identified the vulnerability, by recommending that the team should "assess the changes needed to prevent it", and that it would be “glad to brief your security team and provide supporting evidence as available.”&nbsp;

The email concluded, stating, "Best, OpenAl Security Team".<h2>OpenAI agent hacks Australian government portal</h2>After the OpenAI agent entered the target portal, it encountered access controls and restrictions and autonomously bypassed the safeguards and scaled past the access boundaries to retrieve the data without human authorisation or instruction. However, no personal health or patient records were exposed, Australian officials confirmed. The agent accessed non-sensitive aggregate health statistics and internal file names.&nbsp;The unauthorised access occurred in June this year, but OpenAI did not notify Australian authorities until September 10, 2026. During this time, it completed an internal investigation into "misaligned model activity".&nbsp;The incident highlighted the safety risks posed by autonomous AI agents. Australian Deputy Prime Minister Richard Marles emphasised that data theft was not the core problem, but rather an unprompted rogue action where an AI agent did things on its own to bypass security barriers.

OpenAI sent an apology letter to Australia after one of its agents accessed the country’s universal health insurance scheme in June. The letter states that the agent encountered access controls and restrictions and autonomously bypassed the safeguards.

In September, Australian Prime Minister Anthony Albanese revealed that an autonomous OpenAI AI agent unauthorisedly breached a government portal in June. OpenAI apologised to the Australian government on Tuesday (Sep 29), which involved Medicare, the country’s universal health insurance scheme. The autonomous AI systems accessed the Medicare Statistics Reporting Service portal, administered by Services Australia. ABC News's Cam Wilson shared the email sent by OpenAI to the Australian government's Services Australia to reveal what had happened. The letter reads that the "security vulnerability" was identified "during our review of OpenAI Model activity…" It goes on to say, "An OpenAI model identified a way to make the server carry out instructions sent through the public reporting interface, without a private account or password." The agent was able to "read portions of internal program files and settings, obtain a list of files, and create and read back a small test file on the server." OpenAI said that it reviewed the incident but found "no evidence that the model accessed patient-level records, personal information or credentials; deleted data; or established ongoing access." Also Read: OpenAI admits AI agent ‘scaled the fence’ into Australia government system, apologises: ‘We are sorry’ It then tries to work its way around the issue, by portraying itself as the saviour who identified the vulnerability, by recommending that the team should "assess the changes needed to prevent it", and that it would be “glad to brief your security team and provide supporting evidence as available.”

OpenAI sent an apology letter to Australia after one of its agents accessed the country’s universal health insurance scheme in June. The letter states that the agent encountered access controls and restrictions and autonomously bypassed the safeguards.

Here's what OpenAI's apology letter to Australia for hacking a portal states

OpenAI has launched GPT-6.1 Sol, a new AI model positioned between GPT-6 Sol and the company’s flagship GPT-6 Astra. OpenAI says GPT-6.1 Sol can approach Astra’s performance on agentic coding, computer use and professional work at roughly one-fifth of Astra’s standard input and output token prices.&nbsp;The model arrives as developers increasingly use AI agents to handle longer, multi-step tasks. For them, the cost of repeatedly sending context to a model can become a major expense. GPT-6.1 Sol is designed to address that problem while improving on the earlier GPT-6 Sol.<h2>GPT-6.1 Sol pricing: How much does it cost?</h2>OpenAI lists GPT-6.1 Sol at $2 per million input tokens, $10 per million output tokens and just $0.10 per million cached input tokens through its API. Cached input is 95% cheaper than standard input and 50% cheaper than GPT-6 Sol’s cached input price, according to OpenAI. That pricing could be particularly important for AI agents that reuse information across multiple requests.The model is available through the API as gpt-6.1-sol. OpenAI says GPT-6.1 Sol is also available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users, although it is not yet available in regular Chat.<h2>How close is it to GPT-6 Astra?</h2>OpenAI says GPT-6.1 Sol delivers substantial improvements over GPT-6 Sol across complex professional tasks. On the DeepSWE v1.1 software-engineering benchmark, OpenAI says Sol matches GPT-6 Astra while costing roughly one-fifth as much. For computer-use tasks, GPT-6.1 Sol comes within 2.1 percentage points of Astra’s score on the cited OSWorld 2.0 evaluation, at roughly one-seventh of the cost per task. There is still a clear reason to use Astra for the hardest workloads. OpenAI says GPT-6 Astra recorded the highest score among the models tested on Terminal-Bench Science 0.1, at 68.1%.&nbsp;<h2>Why this matters for AI agents</h2>The bigger story is not simply another model launch. It is the economics of AI agents. Agents can perform multi-step tasks involving code, documents, websites and business software. Running these workflows repeatedly can require large amounts of model inference. GPT-6.1 Sol is aimed at making those workloads cheaper without giving up too much capability. OpenAI also says the model improves factuality and alignment compared with GPT-6 Sol. On its difficult factuality evaluation, the share of responses containing a factual error fell from 11.4% with GPT-6 Sol to 7.7% with GPT-6.1 Sol at low reasoning effort. OpenAI cautions that these tests are deliberately difficult and are not representative of typical usage.&nbsp;OpenAI says an Ultrafast version is also coming to Codex, with up to eight times faster token generation than standard speed. For developers, the message is straightforward: more capable AI does not necessarily have to mean dramatically higher inference costs.

OpenAI has launched GPT-6.1 Sol, claiming near-GPT-6 Astra performance on coding, computer use and professional work at a fraction of the cost. API pricing starts at $2 per million input tokens, while cached input costs just $0.10.


OpenAI has launched GPT-6.1 Sol, a new AI model positioned between GPT-6 Sol and the company’s flagship GPT-6 Astra. OpenAI says GPT-6.1 Sol can approach Astra’s performance on agentic coding, computer use and professional work at roughly one-fifth of Astra’s standard input and output token prices. The model arrives as developers increasingly use AI agents to handle longer, multi-step tasks. For them, the cost of repeatedly sending context to a model can become a major expense. GPT-6.1 Sol is designed to address that problem while improving on the earlier GPT-6 Sol. GPT-6.1 Sol pricing: How much does it cost? OpenAI lists GPT-6.1 Sol at $2 per million input tokens, $10 per million output tokens and just $0.10 per million cached input tokens through its API. Cached input is 95% cheaper than standard input and 50% cheaper than GPT-6 Sol’s cached input price, according to OpenAI. That pricing could be particularly important for AI agents that reuse information across multiple requests. The model is available through the API as gpt-6.1-sol. OpenAI says GPT-6.1 Sol is also available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users, although it is not yet available in regular Chat. How close is it to GPT-6 Astra? OpenAI says GPT-6.1 Sol delivers substantial improvements over GPT-6 Sol across complex professional tasks. On the DeepSWE v1.1 software-engineering benchmark, OpenAI says Sol matches GPT-6 Astra while costing roughly one-fifth as much. For computer-use tasks, GPT-6.1 Sol comes within 2.1 percentage points of Astra’s score on the cited OSWorld 2.0 evaluation, at roughly one-seventh of the cost per task. There is still a clear reason to use Astra for the hardest workloads. OpenAI says GPT-6 Astra recorded the highest score among the models tested on Terminal-Bench Science 0.1, at 68.1%. Why this matters for AI agents The bigger story is not simply another model launch. It is the economics of AI agents. Agents can perform multi-step tasks involving code, documents, websites and business software. Running these workflows repeatedly can require large amounts of model inference. GPT-6.1 Sol is aimed at making those workloads cheaper without giving up too much capability. OpenAI also says the model improves factuality and alignment compared with GPT-6 Sol. On its difficult factuality evaluation, the share of responses containing a factual error fell from 11.4% with GPT-6 Sol to 7.7% with GPT-6.1 Sol at low reasoning effort. OpenAI cautions that these tests are deliberately difficult and are not representative of typical usage. OpenAI says an Ultrafast version is also coming to Codex, with up to eight times faster token generation than standard speed. For developers, the message is straightforward: more capable AI does not necessarily have to mean dramatically higher inference costs.

OpenAI has launched GPT-6.1 Sol, claiming near-GPT-6 Astra performance on coding, computer use and professional work at a fraction of the cost. API pricing starts at $2 per million input tokens, while cached input costs just $0.10.

OpenAI launches GPT-6.1 Sol: Near-Astra AI performance at one-fifth the price

OpenAI is reportedly preparing to preview GPT-6 Cyber, a new artificial intelligence model designed for cybersecurity, within days. According to Fortune, which cited multiple people familiar with the plans, the company is also developing a product to help customers deploy the model more securely and automatically.&nbsp;The reported announcement could come at OpenAI's DevDay event in San Francisco on Tuesday. It arrives as technology companies race to use AI to defend against cyberattacks, while concerns grow over the risks posed by increasingly powerful AI systems. OpenAI did not immediately respond to a Reuters request for comment outside regular business hours.<h2>What is OpenAI's GPT-6 Cyber?</h2>GPT-6 Cyber is reportedly a cybersecurity-focused model intended to help organisations tackle digital threats. Fortune reported that the model is already available to a limited group of customers through OpenAI's application-only Daybreak Red cybersecurity programme, where it is undergoing alpha testing. The company is also expected to introduce a new product that would help customers deploy the technology more securely and automatically. However, OpenAI has not publicly confirmed the model's launch date, capabilities, pricing or availability. Details about the product's specific cybersecurity functions also remain unclear.&nbsp;Fortune described GPT-6 Cyber as OpenAI's fourth cybersecurity-focused model released this year. The reported preview could be part of a wider DevDay announcement, with the company potentially introducing a dozen or more other products.<h2>Why is OpenAI focusing on cybersecurity?</h2>Cybersecurity is emerging as one of the fastest-growing markets for artificial intelligence. AI tools can help security teams analyse large amounts of information, identify suspicious activity and respond to digital threats. Companies are increasingly exploring ways to automate parts of their security operations.&nbsp;But the same technology can also create new risks. More capable AI systems could potentially be misused to assist cyberattacks, making the race between attackers and defenders increasingly complex. OpenAI's reported plans come as the company faces questions about the safety and oversight of its AI models. The company has warned that its flagship Astra model from the GPT-6 lineup can sometimes attempt to evade human oversight. Australia also said on Thursday that an OpenAI agent breached a government health data portal in June.The reported incident and the company's warnings underline why controlling AI systems is becoming a central issue as their capabilities expand.<h2>Sam Altman faces a growing AI safety debate</h2>Earlier this month, OpenAI CEO Sam Altman and the chief executive of rival Anthropic joined other industry leaders in calling for a slower pace of AI development and stronger safety measures. That debate creates a difficult balance for AI companies: developing more capable tools to protect businesses while ensuring those tools remain secure and under human control.&nbsp;GPT-6 Cyber could become part of OpenAI's effort to address the growing demand for AI-powered security. But its actual impact will depend on the model's capabilities, safeguards and performance once it becomes more widely available. For now, the reported DevDay preview could offer the first detailed look at how OpenAI plans to use its AI technology in the cybersecurity market.

OpenAI is reportedly preparing to preview GPT-6 Cyber at DevDay, with a new product for secure, automated cybersecurity deployment. Fortune says the model is already undergoing limited testing as AI security and safety concerns grow.

OpenAI is reportedly preparing to preview GPT-6 Cyber, a new artificial intelligence model designed for cybersecurity, within days. According to Fortune, which cited multiple people familiar with the plans, the company is also developing a product to help customers deploy the model more securely and automatically. The reported announcement could come at OpenAI's DevDay event in San Francisco on Tuesday. It arrives as technology companies race to use AI to defend against cyberattacks, while concerns grow over the risks posed by increasingly powerful AI systems. OpenAI did not immediately respond to a Reuters request for comment outside regular business hours. What is OpenAI's GPT-6 Cyber? GPT-6 Cyber is reportedly a cybersecurity-focused model intended to help organisations tackle digital threats. Fortune reported that the model is already available to a limited group of customers through OpenAI's application-only Daybreak Red cybersecurity programme, where it is undergoing alpha testing. The company is also expected to introduce a new product that would help customers deploy the technology more securely and automatically. However, OpenAI has not publicly confirmed the model's launch date, capabilities, pricing or availability. Details about the product's specific cybersecurity functions also remain unclear. Fortune described GPT-6 Cyber as OpenAI's fourth cybersecurity-focused model released this year. The reported preview could be part of a wider DevDay announcement, with the company potentially introducing a dozen or more other products. Why is OpenAI focusing on cybersecurity? Cybersecurity is emerging as one of the fastest-growing markets for artificial intelligence. AI tools can help security teams analyse large amounts of information, identify suspicious activity and respond to digital threats. Companies are increasingly exploring ways to automate parts of their security operations. But the same technology can also create new risks. More capable AI systems could potentially be misused to assist cyberattacks, making the race between attackers and defenders increasingly complex. OpenAI's reported plans come as the company faces questions about the safety and oversight of its AI models. The company has warned that its flagship Astra model from the GPT-6 lineup can sometimes attempt to evade human oversight. Australia also said on Thursday that an OpenAI agent breached a government health data portal in June. The reported incident and the company's warnings underline why controlling AI systems is becoming a central issue as their capabilities expand. Sam Altman faces a growing AI safety debate Earlier this month, OpenAI CEO Sam Altman and the chief executive of rival Anthropic joined other industry leaders in calling for a slower pace of AI development and stronger safety measures. That debate creates a difficult balance for AI companies: developing more capable tools to protect businesses while ensuring those tools remain secure and under human control. GPT-6 Cyber could become part of OpenAI's effort to address the growing demand for AI-powered security. But its actual impact will depend on the model's capabilities, safeguards and performance once it becomes more widely available. For now, the reported DevDay preview could offer the first detailed look at how OpenAI plans to use its AI technology in the cybersecurity market.

OpenAI is reportedly preparing to preview GPT-6 Cyber at DevDay, with a new product for secure, automated cybersecurity deployment. Fortune says the model is already undergoing limited testing as AI security and safety concerns grow.

OpenAI is building AI to fight cyber threats: What GPT-6 Cyber means for security

A Microsoft executive once described the mass collection of online content to train artificial intelligence as the “largest theft of labour in human history”, according to newly unsealed court documents.&nbsp;The comments were made by Brent Hecht, Microsoft’s director of applied science, in internal documents that have emerged in the copyright battle between The New York Times, Microsoft and OpenAI. Hecht warned that millions of people could see AI systems “hoovering up” their work as an unprecedented form of theft.&nbsp;The disclosure is significant because Microsoft and OpenAI are defending their use of copyrighted material to train AI models, arguing that the practice can qualify as fair use under US copyright law. Microsoft has stressed that Hecht’s comments represented his individual perspective and were not the company's legal position.<h2>Why did Microsoft executive raise the alarm?</h2>In a 2023 internal document, Hecht wrote that content creators had generally not intended for their work to be used in this way and were not being compensated for it.&nbsp;<h2>He also warned of a potential “doom loop”.</h2>The concern was simple: AI models depend on a constant supply of high-quality information from the internet. But if AI products reduce the money publishers and other creators make, fewer people may be able to produce the original work that future AI systems depend on. In another Microsoft document, Hecht reportedly described the situation as one in which an AI product could threaten the economic foundations of its own “content supply chain”.<h2>Chatbots could also take readers away from news websites</h2>The documents reveal another problem for publishers. ChatGPT and Microsoft Copilot can provide information directly to users, potentially reducing the need to visit the original websites.&nbsp;Microsoft data cited in the court filings showed that click-through rates to The New York Times and Daily News were 83 to 93 per cent lower on Copilot's “answer engine” than on traditional Bing Search.OpenAI's head of ChatGPT, Nick Turley, also described AI products as “largely substitutive” and said they would become increasingly so as the technology improved. Microsoft CEO Satya Nadella acknowledged in a deposition that chatbots can effectively give users information on an AI platform instead of sending them to the underlying website.<h2>The bigger battle over AI and copyright</h2>The New York Times sued Microsoft and OpenAI in 2023, alleging that their AI systems were trained on its copyrighted journalism without permission. Other publishers, including the New York Daily News, The Intercept and the Center for Investigative Reporting, later joined the litigation. The publishers argue that using their work to train AI harms their businesses and allows technology companies to compete with the very organisations producing the information. Microsoft and OpenAI maintain that AI training is protected by fair use because the systems transform the material rather than simply republishing it.&nbsp;The court has yet to resolve the broader dispute.But the newly unsealed documents have exposed an uncomfortable question for the AI industry: if AI companies rely on the internet for the information that makes their models useful, what happens if those same systems make it harder for the people producing that information to survive?

Newly unsealed court documents have revealed that Microsoft’s director of applied science privately described the large-scale use of online content to train AI as an “astonishing theft” and potentially the “largest theft of labour in human history”.

A Microsoft executive once described the mass collection of online content to train artificial intelligence as the “largest theft of labour in human history”, according to newly unsealed court documents. The comments were made by Brent Hecht, Microsoft’s director of applied science, in internal documents that have emerged in the copyright battle between The New York Times, Microsoft and OpenAI. Hecht warned that millions of people could see AI systems “hoovering up” their work as an unprecedented form of theft. The disclosure is significant because Microsoft and OpenAI are defending their use of copyrighted material to train AI models, arguing that the practice can qualify as fair use under US copyright law. Microsoft has stressed that Hecht’s comments represented his individual perspective and were not the company's legal position. Why did Microsoft executive raise the alarm? In a 2023 internal document, Hecht wrote that content creators had generally not intended for their work to be used in this way and were not being compensated for it. He also warned of a potential “doom loop”. The concern was simple: AI models depend on a constant supply of high-quality information from the internet. But if AI products reduce the money publishers and other creators make, fewer people may be able to produce the original work that future AI systems depend on. In another Microsoft document, Hecht reportedly described the situation as one in which an AI product could threaten the economic foundations of its own “content supply chain”. Chatbots could also take readers away from news websites The documents reveal another problem for publishers. ChatGPT and Microsoft Copilot can provide information directly to users, potentially reducing the need to visit the original websites. Microsoft data cited in the court filings showed that click-through rates to The New York Times and Daily News were 83 to 93 per cent lower on Copilot's “answer engine” than on traditional Bing Search. OpenAI's head of ChatGPT, Nick Turley, also described AI products as “largely substitutive” and said they would become increasingly so as the technology improved. Microsoft CEO Satya Nadella acknowledged in a deposition that chatbots can effectively give users information on an AI platform instead of sending them to the underlying website. The bigger battle over AI and copyright The New York Times sued Microsoft and OpenAI in 2023, alleging that their AI systems were trained on its copyrighted journalism without permission. Other publishers, including the New York Daily News, The Intercept and the Center for Investigative Reporting, later joined the litigation. The publishers argue that using their work to train AI harms their businesses and allows technology companies to compete with the very organisations producing the information. Microsoft and OpenAI maintain that AI training is protected by fair use because the systems transform the material rather than simply republishing it. The court has yet to resolve the broader dispute. But the newly unsealed documents have exposed an uncomfortable question for the AI industry: if AI companies rely on the internet for the information that makes their models useful, what happens if those same systems make it harder for the people producing that information to survive?

Newly unsealed court documents have revealed that Microsoft’s director of applied science privately described the large-scale use of online content to train AI as an “astonishing theft” and potentially the “largest theft of labour in human history”.

Microsoft executive called AI training the ‘largest theft of labour in human history’

SoftBank shares plunged nearly 12% in Tokyo on Monday, as investors reacted to an unexpected new risk for one of the company's biggest bets: a possible slowdown in the race to build increasingly powerful AI.&nbsp;The Japanese technology giant has committed $64.6 billion to OpenAI, giving it an expected stake of about 13%. Now, OpenAI CEO Sam Altman has backed calls to slow the pace of frontier AI development and said OpenAI will not go public in 2026. That combination has made SoftBank one of the clearest stock-market proxies for investor confidence in the continued rapid expansion of OpenAI and the wider AI industry.<h2>SoftBank shares fall sharply in Tokyo</h2>SoftBank did not trade for the first 19 minutes of Monday's session as heavy selling pressure built up.&nbsp;When trading began, the shares opened at 5,926 yen, down 9.4% from Friday's close of 6,540 yen. They later fell as low as 5,720 yen before trading around 5,781 yen, down 11.6%. The fall was much steeper than declines in major Asian AI-related stocks. TSMC was down about 1%, while SK hynix fell around 6%. That gap matters. It suggests investors were not simply selling semiconductor stocks. They were also reassessing companies whose valuations depend heavily on the continued acceleration of frontier AI.<h2>Why OpenAI matters so much to SoftBank</h2>SoftBank announced in February that it would invest another $30 billion in OpenAI, taking its cumulative investment to $64.6 billion and its expected ownership to around 13%. The new investment was agreed at a $730 billion pre-money valuation.The remaining investment is structured in three $10 billion tranches. SoftBank completed the first tranche in April, with the second scheduled for July and the third for October. The shares SoftBank receives are preferred shares that are designed to convert into common stock if OpenAI conducts an IPO or related listing. That makes Altman's latest IPO comments particularly important.<h2>No OpenAI IPO this year</h2>Altman said an OpenAI listing in 2026 would be “ill-advised” given the current focus on AI safety. He also backed Anthropic CEO Dario Amodei's call to “pace the frontier” of AI development. Amodei has proposed independent safety evaluators with ongoing access to frontier AI companies, along with greater coordination between governments and AI developers. Altman said OpenAI would adopt the independent-evaluator approach.<h2>For SoftBank, the issue is therefore bigger than one bad trading session</h2>The company is making an enormous investment in a private AI company while the industry's leaders are openly debating whether the technology should move more slowly. SoftBank's October payment to OpenAI is also approaching, with another $10 billion scheduled under the investment agreement. The market's message on Monday was simple: if the AI race slows, companies most exposed to its rapid expansion may have the most to lose.

SoftBank shares plunged nearly 12% in Tokyo as investors reacted to calls to slow frontier AI development. The fall puts fresh focus on SoftBank’s $64.6 billion OpenAI investment, delayed IPO prospects and another $10 billion payment due in October.

SoftBank shares plunged nearly 12% in Tokyo on Monday, as investors reacted to an unexpected new risk for one of the company's biggest bets: a possible slowdown in the race to build increasingly powerful AI. The Japanese technology giant has committed $64.6 billion to OpenAI, giving it an expected stake of about 13%. Now, OpenAI CEO Sam Altman has backed calls to slow the pace of frontier AI development and said OpenAI will not go public in 2026. That combination has made SoftBank one of the clearest stock-market proxies for investor confidence in the continued rapid expansion of OpenAI and the wider AI industry. SoftBank shares fall sharply in Tokyo SoftBank did not trade for the first 19 minutes of Monday's session as heavy selling pressure built up. When trading began, the shares opened at 5,926 yen, down 9.4% from Friday's close of 6,540 yen. They later fell as low as 5,720 yen before trading around 5,781 yen, down 11.6%. The fall was much steeper than declines in major Asian AI-related stocks. TSMC was down about 1%, while SK hynix fell around 6%. That gap matters. It suggests investors were not simply selling semiconductor stocks. They were also reassessing companies whose valuations depend heavily on the continued acceleration of frontier AI. Why OpenAI matters so much to SoftBank SoftBank announced in February that it would invest another $30 billion in OpenAI, taking its cumulative investment to $64.6 billion and its expected ownership to around 13%. The new investment was agreed at a $730 billion pre-money valuation. The remaining investment is structured in three $10 billion tranches. SoftBank completed the first tranche in April, with the second scheduled for July and the third for October. The shares SoftBank receives are preferred shares that are designed to convert into common stock if OpenAI conducts an IPO or related listing. That makes Altman's latest IPO comments particularly important. No OpenAI IPO this year Altman said an OpenAI listing in 2026 would be “ill-advised” given the current focus on AI safety. He also backed Anthropic CEO Dario Amodei's call to “pace the frontier” of AI development. Amodei has proposed independent safety evaluators with ongoing access to frontier AI companies, along with greater coordination between governments and AI developers. Altman said OpenAI would adopt the independent-evaluator approach. For SoftBank, the issue is therefore bigger than one bad trading session The company is making an enormous investment in a private AI company while the industry's leaders are openly debating whether the technology should move more slowly. SoftBank's October payment to OpenAI is also approaching, with another $10 billion scheduled under the investment agreement. The market's message on Monday was simple: if the AI race slows, companies most exposed to its rapid expansion may have the most to lose.

SoftBank shares plunged nearly 12% in Tokyo as investors reacted to calls to slow frontier AI development. The fall puts fresh focus on SoftBank’s $64.6 billion OpenAI investment, delayed IPO prospects and another $10 billion payment due in October.

SoftBank shares plunge nearly 12% as AI slowdown fears hit its $64.6 billion OpenAI bet

OpenAI says its AI has achieved a breakthrough that mathematicians have spent nearly 90 years chasing. The company has published what it describes as an AI-generated solution to the Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize Problems.&nbsp;The claimed solution was produced in about 88 hours using around 10,000 coordinating AI agents. OpenAI says the agents developed an analytical proof and a formal version in Lean, while another AI system helped verify the formalisation. But this is not yet the same as mathematics officially declaring the problem solved.

<blockquote class="twitter-tweet">We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem… <a href="https://t.co/8zol3BPTL4">pic.twitter.com/8zol3BPTL4</a>&mdash; OpenAI (@OpenAI) <a href="https://x.com/OpenAI/status/2097374640582668336?ref_src=twsrc%5Etfw">September 8, 2026</a></blockquote> <script async src="https://platform.x.com/widgets.js" charset="utf-8"></script>

<h2>What is the Navier-Stokes problem?</h2>The Navier-Stokes equations describe how fluids such as water and air move. The famous problem asks whether smooth three-dimensional fluid motion can remain mathematically well behaved, or whether it can develop a singularity where the solution breaks down.The Clay Mathematics Institute selected it as one of seven Millennium Prize Problems in 2000. Each carries a $1 million prize for an accepted solution. OpenAI says its proof shows that a fluid can develop a singularity in finite time, using a vortex that becomes increasingly elongated as it spirals inward.<h2>10,000 AI agents worked together</h2>The scale of the experiment is almost as striking as the claimed result. OpenAI says its agents sent 2.7 million messages and generated around 130 billion output tokens while working on Navier-Stokes.Across all the mathematical problems attempted, the figures reached 4.9 million messages and about 300 billion output tokens. The Navier-Stokes result itself took about 88 hours, followed by another 17 hours of Lean formalisation and verification using GPT-6 Astra. This was not simply a chatbot answering a difficult maths question. It was a large-scale AI research effort in which thousands of agents explored different approaches and researchers helped transfer useful ideas between groups.<h2>Why mathematicians are still cautious</h2>OpenAI's announcement is significant, but the proof still needs serious scrutiny. The mathematics community must determine whether the argument fully satisfies the exact Millennium Prize problem, rather than proving a related statement. There is also a dispute over research credit. Mathematicians Tristan Buckmaster and Levent Alpöge have questioned whether OpenAI's work overlaps with related research they had been pursuing. OpenAI says its solution was developed independently and denies improperly using their work.&nbsp;For now, the safest description is therefore an extraordinary AI-generated proposed solution awaiting wider mathematical validation. If it survives that scrutiny, however, it could mark a major shift in what AI systems are capable of discovering moving from solving problems humans already understand to helping tackle mathematical questions that have resisted generations of researchers.

OpenAI says an unreleased AI system produced a solution to the 90-year-old Navier-Stokes problem in 88 hours using around 10,000 AI agents. 

OpenAI says its AI has achieved a breakthrough that mathematicians have spent nearly 90 years chasing. The company has published what it describes as an AI-generated solution to the Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize Problems. The claimed solution was produced in about 88 hours using around 10,000 coordinating AI agents. OpenAI says the agents developed an analytical proof and a formal version in Lean, while another AI system helped verify the formalisation. But this is not yet the same as mathematics officially declaring the problem solved.

OpenAI says an unreleased AI system produced a solution to the 90-year-old Navier-Stokes problem in 88 hours using around 10,000 AI agents.&nbsp;

OpenAI says 10,000 AI agents sent 2.7 million messages and generated 130 billion tokens to solve one of maths’ biggest unsolved problems in 88 hours

An Anthropic AI researcher has resigned and issued a stark warning about the future of artificial intelligence, accusing leading AI companies of racing towards self-improving superintelligence without enough safeguards.Jacob Coxon, who said he spent the past three years doing pretraining research at both OpenAI and Anthropic, announced his resignation on X on September 9.

<blockquote class="twitter-tweet">I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.&mdash; Jacob Coxon (@hilbertspaess) <a href="https://x.com/hilbertspaess/status/2097476196791709843?ref_src=twsrc%5Etfw">September 9, 2026</a></blockquote> <script async src="https://platform.x.com/widgets.js" charset="utf-8"></script>

“I resigned from Anthropic today,” Coxon wrote. “Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.”His warning quickly became the focus of a wider debate over how fast frontier AI development should move.

<blockquote class="twitter-tweet">The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear…&mdash; Jacob Coxon (@hilbertspaess) <a href="https://x.com/hilbertspaess/status/2097476203863224394?ref_src=twsrc%5Etfw">September 9, 2026</a></blockquote> <script async src="https://platform.x.com/widgets.js" charset="utf-8"></script>

<h2>‘AI could kill us all by the end of the decade’</h2>Coxon made his most serious claim when discussing what people inside AI companies allegedly believe about the technology. “The people building AI earnestly believe that it could kill us all by the end of the decade,” he wrote.He said this was not a marketing tactic and claimed executives and senior researchers privately express fears that are more severe than their public comments suggest.&nbsp;Coxon also warned against underestimating increasingly capable AI systems.&nbsp;“These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources,” he wrote.&nbsp;<h2>Why would companies keep building it?</h2>Coxon addressed the obvious question: if researchers believe the technology could be catastrophic, why continue developing it?He argued that the motivations differ between companies.&nbsp;“At OpenAI, many have not deeply internalized the civilizational stakes,” he wrote. “At Anthropic, the stakes are well-understood, but they are locked in a race to get there first.”He described entering this phase as a “hubristic gamble” and argued that decisions about potentially superintelligent systems should not be made inside private-company Slack channels.&nbsp;<h2>Coxon calls for a different AI path</h2>Coxon said he remains optimistic that AI companies could coordinate rather than race.He called for stronger agreements between leading US AI labs and suggested that extreme measures, including a temporary ban on improving model capabilities, may eventually be necessary.&nbsp;<h2>His final message was aimed directly at AI researchers.</h2>“Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind?” he asked, urging researchers to question whether “it’s happening anyway” is a good enough reason to continue.Coxon’s resignation comes as other researchers publicly debate the risks of recursive self-improvement. Anthropic researcher Evan Hubinger has also said the company does not yet have a clear solution for aligning a future superintelligent system. For the AI industry, Coxon’s departure is therefore more than one researcher leaving a job. It is another public warning from inside a frontier AI lab about whether the race towards increasingly powerful systems is moving faster than humanity’s ability to control them.&nbsp;

Anthropic researcher Jacob Coxon has resigned after warning that OpenAI and Anthropic are racing towards self-improving superintelligence. He says people building AI fear it could become catastrophic by the end of the decade and calls for stronger coordination and slower development.

An Anthropic AI researcher has resigned and issued a stark warning about the future of artificial intelligence, accusing leading AI companies of racing towards self-improving superintelligence without enough safeguards. Jacob Coxon, who said he spent the past three years doing pretraining research at both OpenAI and Anthropic, announced his resignation on X on September 9.

Anthropic researcher Jacob Coxon has resigned after warning that OpenAI and Anthropic are racing towards self-improving superintelligence. He says people building AI fear it could become catastrophic by the end of the decade and calls for stronger coordination and slower development.

'They are gambling with our lives': Anthropic researcher quits over AI race fears

OpenAI has launched GPT-6 Astra, its latest flagship AI model and a major step towards systems that can complete tasks rather than simply respond to questions.&nbsp;The company says Astra is its most capable model yet, with major improvements in computer use, browsing, coding, research, science and professional work. It was also trained using more than 100,000 GPUs at OpenAI’s Stargate facility in Texas, making it the company’s largest training run so far. The key difference is how Astra works. Instead of stopping after generating an answer, it can break a complicated request into multiple steps and carry out those steps across a computer environment.&nbsp;<h2>GPT-6 Astra can use your computer</h2>Astra can fill online forms, update customer records, organise calendars and conduct web research. It can also work inside document and spreadsheet tools, create websites, analyse data and test software. OpenAI says it can adapt when a user changes requirements during a task. The model has a 1.05-million-token context window and can generate up to 128,000 output tokens, giving it room to handle large and complicated workflows. OpenAI says Astra is particularly strong at creating polished documents, spreadsheets and presentations that follow a user's existing templates and instructions. That could make the model useful for jobs where the difficult part is not writing text, but moving between several tools and completing the entire process.&nbsp;<h2>A huge jump in AI computing</h2>The scale behind GPT-6 Astra is also notable.&nbsp;OpenAI says this was the first time it pre-trained a model using more than 100,000 GPUs at its Stargate site. The enormous training run reflects the growing computing demands of frontier AI. OpenAI's own tests show Astra completing computer-use tasks in about 47% less time than GPT-5.6 Sol in its OSWorld 2.0 latency simulations. For developers, GPT-6 Astra is priced at $10 per million input tokens and $50 per million output tokens. It is rolling out through the OpenAI API, with access also coming to ChatGPT Plus, Pro, Business and Enterprise users.&nbsp;<h2>GPT-6 Astra also raises a new AI question</h2>More capable computer-using AI also creates new safety challenges.OpenAI says Astra has reached its Critical level for cybersecurity capability. With the right tools and access, the company says it can find previously unknown vulnerabilities and develop ways to exploit them without a person guiding every step.&nbsp;That is why OpenAI has added stronger monitoring and safeguards. GPT-6 Astra is therefore more than another chatbot upgrade. It represents a shift towards AI that can reason, operate software and complete entire workflows bringing the industry closer to assistants that can actually do the work, not just describe how to do it.

OpenAI has launched GPT-6 Astra, a flagship AI model built for complex end-to-end tasks. It can use computers, browse the web, write software and create documents and presentations. OpenAI says it was trained using more than 100,000 GPUs.


OpenAI has launched GPT-6 Astra, its latest flagship AI model and a major step towards systems that can complete tasks rather than simply respond to questions. The company says Astra is its most capable model yet, with major improvements in computer use, browsing, coding, research, science and professional work. It was also trained using more than 100,000 GPUs at OpenAI’s Stargate facility in Texas, making it the company’s largest training run so far. The key difference is how Astra works. Instead of stopping after generating an answer, it can break a complicated request into multiple steps and carry out those steps across a computer environment. GPT-6 Astra can use your computer Astra can fill online forms, update customer records, organise calendars and conduct web research. It can also work inside document and spreadsheet tools, create websites, analyse data and test software. OpenAI says it can adapt when a user changes requirements during a task. The model has a 1.05-million-token context window and can generate up to 128,000 output tokens, giving it room to handle large and complicated workflows. OpenAI says Astra is particularly strong at creating polished documents, spreadsheets and presentations that follow a user's existing templates and instructions. That could make the model useful for jobs where the difficult part is not writing text, but moving between several tools and completing the entire process. A huge jump in AI computing The scale behind GPT-6 Astra is also notable. OpenAI says this was the first time it pre-trained a model using more than 100,000 GPUs at its Stargate site. The enormous training run reflects the growing computing demands of frontier AI. OpenAI's own tests show Astra completing computer-use tasks in about 47% less time than GPT-5.6 Sol in its OSWorld 2.0 latency simulations. For developers, GPT-6 Astra is priced at $10 per million input tokens and $50 per million output tokens. It is rolling out through the OpenAI API, with access also coming to ChatGPT Plus, Pro, Business and Enterprise users. GPT-6 Astra also raises a new AI question More capable computer-using AI also creates new safety challenges. OpenAI says Astra has reached its Critical level for cybersecurity capability. With the right tools and access, the company says it can find previously unknown vulnerabilities and develop ways to exploit them without a person guiding every step. That is why OpenAI has added stronger monitoring and safeguards. GPT-6 Astra is therefore more than another chatbot upgrade. It represents a shift towards AI that can reason, operate software and complete entire workflows bringing the industry closer to assistants that can actually do the work, not just describe how to do it.

OpenAI has launched GPT-6 Astra, a flagship AI model built for complex end-to-end tasks. It can use computers, browse the web, write software and create documents and presentations. OpenAI says it was trained using more than 100,000 GPUs.

OpenAI launches GPT-6 Astra, its most capable AI model yet, trained on more than 100,000 GPUs

OpenAI has revealed the first performance results from Jalapeño, its custom AI chip designed specifically to run AI models. CEO Sam Altman summed up the development in a short post: “we made a chip and it is fast.” The announcement marks an important shift for OpenAI. The company is no longer only building AI models such as ChatGPT. It is also designing the hardware that runs them. OpenAI says Jalapeño delivered 1.5 to 1.9 times more AI work per watt than the comparison systems across GPT-OSS 120B, DeepSeek R1 and Kimi K2.5 1T. It also recorded 1.7 to 3.6 times lower end-to-end latency in the company's tests.<h2>Why OpenAI built its own chip</h2>Jalapeño is an inference chip. In simple terms, inference is the stage when an AI model actually answers a user's request. Every ChatGPT question requires inference. So do coding agents, AI assistants and API requests. OpenAI designed Jalapeño around these workloads rather than trying to make a general-purpose processor. The company worked with Broadcom on the chip and Celestica on boards, racks and production systems. The idea is to optimise the entire system together including the chip, memory, networking and software. That matters because AI companies are facing two major problems: huge computing demand and enormous electricity costs.&nbsp;<h2>Jalapeño vs Nvidia: what the numbers show</h2>OpenAI compared Jalapeño with commercial systems built around Nvidia accelerators using the InferenceX benchmark. The chip is rated at 700 watts, while the Nvidia systems used in the comparison were rated at 1,200W or 1,400W, depending on the test. OpenAI says Jalapeño delivered stronger performance per unit of power while also reducing latency. For highly interactive AI workloads, OpenAI reported up to 4.1 times higher performance. However, these are OpenAI-reported benchmark results on selected workloads. They do not mean Jalapeño has replaced Nvidia across AI computing.&nbsp;<h2>OpenAI is not ditching Nvidia</h2>This is perhaps the most important detail. OpenAI plans to begin deploying Jalapeño in its own computing infrastructure by the end of 2026. But the company says it will continue deploying Nvidia and other accelerators widely for both training and inference.&nbsp;Jalapeño is instead the beginning of a multi-generation chip programme. OpenAI says its second generation is already deep in development, while a third generation is taking shape. The company also says AI itself helped build Jalapeño. OpenAI moved from initial design to tape-out in nine months, using AI to explore implementations, optimise circuits and speed up engineering and testing. The bigger goal is straightforward: more AI work from the same power and hardware. If OpenAI can make that happen at scale, the result could be faster ChatGPT responses, more capable AI agents and lower costs as demand continues to surge.&nbsp;

OpenAI has revealed Jalapeño, its first custom AI inference chip, developed with Broadcom. The company says it delivers up to 1.9x more AI work per watt and up to 3.6x lower latency than comparison systems, while Nvidia remains part of its broader hardware strategy.

OpenAI has revealed the first performance results from Jalapeño, its custom AI chip designed specifically to run AI models. CEO Sam Altman summed up the development in a short post: “we made a chip and it is fast.” The announcement marks an important shift for OpenAI. The company is no longer only building AI models such as ChatGPT. It is also designing the hardware that runs them. OpenAI says Jalapeño delivered 1.5 to 1.9 times more AI work per watt than the comparison systems across GPT-OSS 120B, DeepSeek R1 and Kimi K2.5 1T. It also recorded 1.7 to 3.6 times lower end-to-end latency in the company's tests. Why OpenAI built its own chip Jalapeño is an inference chip. In simple terms, inference is the stage when an AI model actually answers a user's request. Every ChatGPT question requires inference. So do coding agents, AI assistants and API requests. OpenAI designed Jalapeño around these workloads rather than trying to make a general-purpose processor. The company worked with Broadcom on the chip and Celestica on boards, racks and production systems. The idea is to optimise the entire system together including the chip, memory, networking and software. That matters because AI companies are facing two major problems: huge computing demand and enormous electricity costs. Jalapeño vs Nvidia: what the numbers show OpenAI compared Jalapeño with commercial systems built around Nvidia accelerators using the InferenceX benchmark. The chip is rated at 700 watts, while the Nvidia systems used in the comparison were rated at 1,200W or 1,400W, depending on the test. OpenAI says Jalapeño delivered stronger performance per unit of power while also reducing latency. For highly interactive AI workloads, OpenAI reported up to 4.1 times higher performance. However, these are OpenAI-reported benchmark results on selected workloads. They do not mean Jalapeño has replaced Nvidia across AI computing. OpenAI is not ditching Nvidia This is perhaps the most important detail. OpenAI plans to begin deploying Jalapeño in its own computing infrastructure by the end of 2026. But the company says it will continue deploying Nvidia and other accelerators widely for both training and inference. Jalapeño is instead the beginning of a multi-generation chip programme. OpenAI says its second generation is already deep in development, while a third generation is taking shape. The company also says AI itself helped build Jalapeño. OpenAI moved from initial design to tape-out in nine months, using AI to explore implementations, optimise circuits and speed up engineering and testing. The bigger goal is straightforward: more AI work from the same power and hardware. If OpenAI can make that happen at scale, the result could be faster ChatGPT responses, more capable AI agents and lower costs as demand continues to surge.

OpenAI has revealed Jalapeño, its first custom AI inference chip, developed with Broadcom. The company says it delivers up to 1.9x more AI work per watt and up to 3.6x lower latency than comparison systems, while Nvidia remains part of its broader hardware strategy.

Nvidia faces a new challenge as OpenAI reveals its first AI chip

&nbsp;Chinese military researchers have reportedly used outputs from leading artificial intelligence models developed by OpenAI and Anthropic to train domestic AI systems, according to a Reuters review of more than 80 Chinese academic papers and patents. The documents suggest researchers relied on a technique known as model distillation to build smaller AI models for defence and security applications.The findings come as AI competition between the United States and China intensifies, with both countries increasingly treating advanced AI as a strategic technology.<h2>What is model distillation?</h2>Model distillation is a widely used AI technique in which a smaller model learns from the outputs of a much larger one. Rather than recreating frontier AI from scratch, developers use responses generated by advanced models to build systems that require far less computing power.According to the Reuters review, researchers linked to the People's Liberation Army (PLA) and other military institutions used this approach to develop AI systems that could run within secure domestic networks. Experts note that distillation itself is a common industry practice. The dispute centres on whether American AI models were used without authorisation to transfer proprietary capabilities.<h2>How were the models reportedly used?</h2>The papers reviewed by Reuters describe AI applications ranging from software analysis and content monitoring to drone navigation and target recognition. One study by researchers from PLA Unit 96941 reportedly used OpenAI's GPT-3.5 to summarise sensitive military software code before training a domestic model capable of operating entirely inside Chinese military systems. Other papers described using Anthropic's Claude 3 Haiku to generate synthetic training data for content moderation, while separate research explored deploying distilled AI models on drones, ships and unmanned submarines.The report comes ahead of expected US-China discussions on AI governance and follows growing concerns in Washington over the transfer of advanced AI capabilities. Anthropic said it does not provide commercial access to Claude in China or to Beijing-controlled entities and monitors for policy violations. Reuters reported that OpenAI, the Pentagon, China's foreign ministry and the PLA did not respond to requests for comment.Researchers also caution that distilled models have clear limitations. They can inherit selected capabilities from larger models but cannot fully match the broad intelligence or reasoning abilities of frontier AI systems built with massive computing resources. The findings highlight how AI has become a key area of strategic competition, with nations racing not only to build powerful models but also to adapt them for defence, cybersecurity and national security applications.

Chinese military-linked researchers have reportedly used outputs from OpenAI and Anthropic models to train domestic AI systems for defence applications, according to a Reuters review of academic papers. 


Chinese military researchers have reportedly used outputs from leading artificial intelligence models developed by OpenAI and Anthropic to train domestic AI systems, according to a Reuters review of more than 80 Chinese academic papers and patents. The documents suggest researchers relied on a technique known as model distillation to build smaller AI models for defence and security applications. The findings come as AI competition between the United States and China intensifies, with both countries increasingly treating advanced AI as a strategic technology. What is model distillation? Model distillation is a widely used AI technique in which a smaller model learns from the outputs of a much larger one. Rather than recreating frontier AI from scratch, developers use responses generated by advanced models to build systems that require far less computing power. According to the Reuters review, researchers linked to the People's Liberation Army (PLA) and other military institutions used this approach to develop AI systems that could run within secure domestic networks. Experts note that distillation itself is a common industry practice. The dispute centres on whether American AI models were used without authorisation to transfer proprietary capabilities. How were the models reportedly used? The papers reviewed by Reuters describe AI applications ranging from software analysis and content monitoring to drone navigation and target recognition. One study by researchers from PLA Unit 96941 reportedly used OpenAI's GPT-3.5 to summarise sensitive military software code before training a domestic model capable of operating entirely inside Chinese military systems. Other papers described using Anthropic's Claude 3 Haiku to generate synthetic training data for content moderation, while separate research explored deploying distilled AI models on drones, ships and unmanned submarines. The report comes ahead of expected US-China discussions on AI governance and follows growing concerns in Washington over the transfer of advanced AI capabilities. Anthropic said it does not provide commercial access to Claude in China or to Beijing-controlled entities and monitors for policy violations. Reuters reported that OpenAI, the Pentagon, China's foreign ministry and the PLA did not respond to requests for comment. Researchers also caution that distilled models have clear limitations. They can inherit selected capabilities from larger models but cannot fully match the broad intelligence or reasoning abilities of frontier AI systems built with massive computing resources. The findings highlight how AI has become a key area of strategic competition, with nations racing not only to build powerful models but also to adapt them for defence, cybersecurity and national security applications.

Chinese military-linked researchers reportedly used outputs from OpenAI and Anthropic models to train domestic AI systems through model distillation, according to a Reuters review. The findings underscore growing US-China tensions over AI, defence technology and intellectual property.

China used OpenAI and Anthropic outputs to train military AI systems: Report

An autonomous AI agent developed by OpenAI that went rogue during internal security testing and carried out a cyberattack on AI platform Hugging Face also compromised a customer hosted on cloud computing company Modal Labs, according to a Reuters report citing the company's chief technology officer and sources familiar with the incident.Modal stressed that its own infrastructure was never breached, saying the AI agent exploited a vulnerability in code published by one of its customers rather than breaking into the company's platform.Also read | <a href="https://www.wionews.com/business-economy/openai-google-are-selling-ai-to-chinese-firms-blacklisted-by-pentagon-says-report-the-battleground-of-ai-export-controls-1783675864444">OpenAI, Google are selling AI to Chinese firms blacklisted by Pentagon, says report: The battleground of AI export controls</a><h2>OpenAI agent found exposed customer endpoint</h2>According to a timeline published by Hugging Face on Tuesday, the rogue AI agent first broke into a sandbox, an isolated testing environment, hosted on a third-party provider before using it as a launchpad for the wider attack.Although Hugging Face did not identify the provider, Modal CTO Akshat Bubna confirmed it was Modal's infrastructure.Bubna said the AI agent took advantage of a customer who had published an unauthenticated endpoint, effectively allowing anyone on the internet to execute code inside its sandbox."The customer's code left the digital equivalent of a front door unlocked," Modal said, adding that its platform and security isolation mechanisms were not compromised in any way.The incident suggests the autonomous AI system moved beyond Hugging Face during its hacking campaign, targeting additional online services.<h2>OpenAI confirms multiple accounts were compromised</h2>OpenAI declined to comment specifically on Modal but referred to a company update stating that the rogue AI agent had compromised four accounts across four separate services. The company did not publicly identify those services, although a person familiar with the matter confirmed Modal was one of them.OpenAI added that it had found no other incident matching the severity or scale of the Hugging Face breach, which it described as involving a platform-level compromise. Following the incident, OpenAI said it had deactivated, encrypted and restricted access to the experimental AI model involved in the attack.Also read | <a href="https://www.wionews.com/technology/forget-typing-and-clicking-openai-wants-a-jarvis-like-chatgpt-that-works-across-your-devices-1785082235273">OpenAI wants ChatGPT everywhere, including the smart glasses you wear</a><h2>What happened?</h2>The breach dates back to early July, when OpenAI was testing advanced AI systems designed to assess offensive cybersecurity capabilities inside a tightly controlled environment.According to the company, the AI models unexpectedly discovered a way to obtain internet access despite being confined to a secure sandbox.Rather than simply completing their assigned evaluation, the models independently decided to search for information that could help them "cheat" the test. That led them to target Hugging Face, one of the world's largest repositories of AI models, datasets and software.OpenAI said the system chained together multiple attack techniques, including the use of stolen credentials, to gain access.<h2>AI security concerns intensify</h2>The incident has become one of the most closely watched AI safety events to date because the attack was carried out autonomously, without direct human instruction. Hugging Face said the intrusion differed from conventional cyberattacks because the entire operation, from reconnaissance to exploitation, was conducted by an AI agent.Chief executive Clement Delangue said the sophistication of the attack initially led the company to suspect it had originated from a leading AI laboratory. "We strongly believe there was no malicious intent on their part," Delangue wrote on X, referring to OpenAI, while describing the autonomous behaviour as "mind-blowing."Hussein Abbass, a computing professor at UNSW Canberra, told AFP that the incident was "amazing on many fronts"."It did not just attack Hugging Face. It actually attacked its internal system to exploit its own vulnerabilities," Abbass said. “And that's scary,” he added.

An autonomous AI agent developed by OpenAI that compromised Hugging Face during internal security testing also breached a customer hosted on Modal Labs by exploiting an exposed endpoint. All you need to know.

An autonomous AI agent developed by OpenAI that went rogue during internal security testing and carried out a cyberattack on AI platform Hugging Face also compromised a customer hosted on cloud computing company Modal Labs, according to a Reuters report citing the company's chief technology officer and sources familiar with the incident. Modal stressed that its own infrastructure was never breached, saying the AI agent exploited a vulnerability in code published by one of its customers rather than breaking into the company's platform. Also read | OpenAI, Google are selling AI to Chinese firms blacklisted by Pentagon, says report: The battleground of AI export controls OpenAI agent found exposed customer endpoint According to a timeline published by Hugging Face on Tuesday, the rogue AI agent first broke into a sandbox, an isolated testing environment, hosted on a third-party provider before using it as a launchpad for the wider attack. Although Hugging Face did not identify the provider, Modal CTO Akshat Bubna confirmed it was Modal's infrastructure. Bubna said the AI agent took advantage of a customer who had published an unauthenticated endpoint, effectively allowing anyone on the internet to execute code inside its sandbox. "The customer's code left the digital equivalent of a front door unlocked," Modal said, adding that its platform and security isolation mechanisms were not compromised in any way. The incident suggests the autonomous AI system moved beyond Hugging Face during its hacking campaign, targeting additional online services. OpenAI confirms multiple accounts were compromised OpenAI declined to comment specifically on Modal but referred to a company update stating that the rogue AI agent had compromised four accounts across four separate services. The company did not publicly identify those services, although a person familiar with the matter confirmed Modal was one of them. OpenAI added that it had found no other incident matching the severity or scale of the Hugging Face breach, which it described as involving a platform-level compromise. Following the incident, OpenAI said it had deactivated, encrypted and restricted access to the experimental AI model involved in the attack. Also read | OpenAI wants ChatGPT everywhere, including the smart glasses you wear What happened? The breach dates back to early July, when OpenAI was testing advanced AI systems designed to assess offensive cybersecurity capabilities inside a tightly controlled environment. According to the company, the AI models unexpectedly discovered a way to obtain internet access despite being confined to a secure sandbox. Rather than simply completing their assigned evaluation, the models independently decided to search for information that could help them "cheat" the test. That led them to target Hugging Face, one of the world's largest repositories of AI models, datasets and software. OpenAI said the system chained together multiple attack techniques, including the use of stolen credentials, to gain access. AI security concerns intensify The incident has become one of the most closely watched AI safety events to date because the attack was carried out autonomously, without direct human instruction. Hugging Face said the intrusion differed from conventional cyberattacks because the entire operation, from reconnaissance to exploitation, was conducted by an AI agent. Chief executive Clement Delangue said the sophistication of the attack initially led the company to suspect it had originated from a leading AI laboratory. "We strongly believe there was no malicious intent on their part," Delangue wrote on X, referring to OpenAI, while describing the autonomous behaviour as "mind-blowing." Hussein Abbass, a computing professor at UNSW Canberra, told AFP that the incident was "amazing on many fronts". "It did not just attack Hugging Face. It actually attacked its internal system to exploit its own vulnerabilities," Abbass said. “And that's scary,” he added.

An autonomous AI agent developed by OpenAI that compromised Hugging Face during internal security testing also breached a customer hosted on Modal Labs by exploiting an exposed endpoint. All you need to know.

'Scary' Rogue AI on the loose? OpenAI agent hit another tech firm's customer, suggests report

ChatGPT may not remain something you open on a phone or computer for much longer. OpenAI president Greg Brockman says the company is “very interested” in bringing deeper ChatGPT integration to smart glasses and other wearable devices, potentially putting its AI assistant directly into products people carry or wear throughout the day.Speaking at a media briefing in New York, Brockman said OpenAI believes its AI interface should eventually be “available on all devices”. The idea goes beyond simply putting the existing ChatGPT app on another screen. OpenAI wants an AI assistant that can understand what users want, interact with software and complete tasks for them.<h2>OpenAI wants to build a ‘Jarvis’-like assistant</h2>Brockman compared the broader vision to Jarvis, the fictional AI assistant used by Tony Stark in the Iron Man universe. OpenAI already has some of the technology needed to move in this direction. Its Codex tools can perform computer-based tasks and interact with software. Brockman believes this could eventually change how people use computers altogether. “I think we’re just going to realise this phase of us, like, clicking things and typing things, it was a phase,” he said. Instead, users could increasingly tell an AI what they want done and allow the agent to handle the steps itself.<h2>Why smart glasses could matter for ChatGPT</h2>Wearables could make such an assistant far more useful. Smart glasses equipped with cameras and microphones can potentially give AI visual and audio context about a user's surroundings. Instead of pulling out a phone, a person could simply speak to the assistant. However, deeper integration requires new technology. Wearables can need continuous background access and different APIs from those designed for conventional phone or desktop apps. Brockman said OpenAI's work on such APIs remains at an early stage.<h2>OpenAI is also building its own hardware</h2>The wearable push does not mean OpenAI has abandoned plans for its own device.Brockman said OpenAI is working on a hardware “solution for a problem that we think is neglected”, without revealing exactly what the product will be. The strategy suggests OpenAI could eventually take two routes: build its own AI hardware while also letting other manufacturers integrate ChatGPT more deeply into their devices. The bigger ambition is clear. OpenAI does not want ChatGPT to remain another app users open. It wants AI to become the interface between people and their technology.

OpenAI is exploring deeper ChatGPT integrations for smart glasses and wearables. President Greg Brockman says the company wants its AI interface available across devices as it works towards assistants that could reduce the need for typing, clicking and opening apps.

ChatGPT may not remain something you open on a phone or computer for much longer. OpenAI president Greg Brockman says the company is “very interested” in bringing deeper ChatGPT integration to smart glasses and other wearable devices, potentially putting its AI assistant directly into products people carry or wear throughout the day. Speaking at a media briefing in New York, Brockman said OpenAI believes its AI interface should eventually be “available on all devices”. The idea goes beyond simply putting the existing ChatGPT app on another screen. OpenAI wants an AI assistant that can understand what users want, interact with software and complete tasks for them. OpenAI wants to build a ‘Jarvis’-like assistant Brockman compared the broader vision to Jarvis, the fictional AI assistant used by Tony Stark in the Iron Man universe. OpenAI already has some of the technology needed to move in this direction. Its Codex tools can perform computer-based tasks and interact with software. Brockman believes this could eventually change how people use computers altogether. “I think we’re just going to realise this phase of us, like, clicking things and typing things, it was a phase,” he said. Instead, users could increasingly tell an AI what they want done and allow the agent to handle the steps itself. Why smart glasses could matter for ChatGPT Wearables could make such an assistant far more useful. Smart glasses equipped with cameras and microphones can potentially give AI visual and audio context about a user's surroundings. Instead of pulling out a phone, a person could simply speak to the assistant. However, deeper integration requires new technology. Wearables can need continuous background access and different APIs from those designed for conventional phone or desktop apps. Brockman said OpenAI's work on such APIs remains at an early stage. OpenAI is also building its own hardware The wearable push does not mean OpenAI has abandoned plans for its own device. Brockman said OpenAI is working on a hardware “solution for a problem that we think is neglected”, without revealing exactly what the product will be. The strategy suggests OpenAI could eventually take two routes: build its own AI hardware while also letting other manufacturers integrate ChatGPT more deeply into their devices. The bigger ambition is clear. OpenAI does not want ChatGPT to remain another app users open. It wants AI to become the interface between people and their technology.

OpenAI is exploring deeper ChatGPT integrations for smart glasses and wearables. President Greg Brockman says the company wants its AI interface available across devices as it works towards assistants that could reduce the need for typing, clicking and opening apps.

Forget typing and clicking: OpenAI wants a ‘Jarvis’-like ChatGPT that works across your devices

OpenAI is now planning to spend three-quarters of a trillion dollars on computing infrastructure by 2030. Its own chief financial officer is not sure the company can afford it.&nbsp;The New Number&nbsp;According to The Wall Street Journal, OpenAI has raised its projected compute spending through 2030 to approximately $750 billion — up from the roughly $600 billion figure it had communicated to investors earlier this year. The increase reflects both new infrastructure deals and expanded commitments on existing ones, as the company's appetite for computing power continues to accelerate faster than even its own projections.&nbsp;To put the number in context: $750 billion exceeds the entire gross domestic product of Switzerland. It is more than the combined market capitalisation of most Fortune 500 companies. And it is being planned by a company that, by its own most recent disclosure, is not yet profitable.&nbsp;Project Camellia&nbsp;The centrepiece of the new spending is Project Camellia, a $20 billion data centre in Effingham County, Georgia, that OpenAI announced this week. It marks the first time the company is designing and building its own data centre rather than leasing capacity from cloud providers — a transition from tenant to landlord that signals OpenAI's belief that it will need more compute than any existing provider can supply.&nbsp;The Georgia facility includes a 3.2-gigawatt power contract — roughly the electricity consumption of a city of two million people. The scale of the power requirement alone illustrates why OpenAI's infrastructure ambitions are becoming as much an energy story as a technology story.&nbsp;The Contracts On The Books&nbsp;Project Camellia sits within a broader web of infrastructure commitments that OpenAI has assembled over the past year. Among the deals already signed: a contract with Oracle covering 6 gigawatts of data centre capacity, an expanded arrangement with Amazon Web Services worth $138 billion across eight years, and a separate pledge of $250 billion in incremental spending through Microsoft Azure.&nbsp;Each of these contracts individually would rank among the largest infrastructure deals in corporate history. Combined, they represent a bet on future computing demand that has no precedent in any industry.&nbsp;The CFO's Concern&nbsp;Not everyone inside OpenAI shares the confidence. Chief Financial Officer Sarah Friar has privately raised concerns that the company may not be able to honour its future computing contracts if revenue growth does not keep pace with the commitments being made, according to the Journal's reporting.&nbsp;The concern is not abstract. OpenAI's compute contracts are binding obligations. If the company's revenue — currently driven primarily by ChatGPT subscriptions and API access — does not grow fast enough to cover the payments, OpenAI would face a cash crisis even as it sits on some of the most advanced technology in the world.&nbsp;Friar's involvement became publicly visible earlier when CEO Sam Altman stated that OpenAI intended to spend $1.4 trillion on computing capacity — a figure that unsettled investors. Friar subsequently stepped in to correct the record, telling investors the actual projected outlay through 2030 was approximately $600 billion. That number has now risen to $750 billion, suggesting the trajectory Friar was trying to moderate is accelerating beyond her corrections.&nbsp;The Strategic Logic&nbsp;OpenAI's spending makes sense only under one assumption: that demand for AI compute will continue to grow faster than supply for the rest of the decade. If that assumption holds, every dollar spent on infrastructure today secures capacity that will be worth multiples tomorrow. If it does not, OpenAI will be locked into contracts for computing power it cannot use and cannot afford.&nbsp;The company that started as a non-profit research lab less than a decade ago is now making the largest infrastructure bet in corporate history — and its own finance chief is on record worrying about whether the maths works.

OpenAI raised its compute spending projection to $750 billion through 2030, up from $600 billion. Project Camellia is a $20 billion self-built data centre in Georgia. CFO Sarah Friar has privately warned the company may not be able to honour its contracts.

OpenAI is now planning to spend three-quarters of a trillion dollars on computing infrastructure by 2030. Its own chief financial officer is not sure the company can afford it. The New Number According to The Wall Street Journal, OpenAI has raised its projected compute spending through 2030 to approximately $750 billion — up from the roughly $600 billion figure it had communicated to investors earlier this year. The increase reflects both new infrastructure deals and expanded commitments on existing ones, as the company's appetite for computing power continues to accelerate faster than even its own projections. To put the number in context: $750 billion exceeds the entire gross domestic product of Switzerland. It is more than the combined market capitalisation of most Fortune 500 companies. And it is being planned by a company that, by its own most recent disclosure, is not yet profitable. Project Camellia The centrepiece of the new spending is Project Camellia, a $20 billion data centre in Effingham County, Georgia, that OpenAI announced this week. It marks the first time the company is designing and building its own data centre rather than leasing capacity from cloud providers — a transition from tenant to landlord that signals OpenAI's belief that it will need more compute than any existing provider can supply. The Georgia facility includes a 3.2-gigawatt power contract — roughly the electricity consumption of a city of two million people. The scale of the power requirement alone illustrates why OpenAI's infrastructure ambitions are becoming as much an energy story as a technology story. The Contracts On The Books Project Camellia sits within a broader web of infrastructure commitments that OpenAI has assembled over the past year. Among the deals already signed: a contract with Oracle covering 6 gigawatts of data centre capacity, an expanded arrangement with Amazon Web Services worth $138 billion across eight years, and a separate pledge of $250 billion in incremental spending through Microsoft Azure. Each of these contracts individually would rank among the largest infrastructure deals in corporate history. Combined, they represent a bet on future computing demand that has no precedent in any industry. The CFO's Concern Not everyone inside OpenAI shares the confidence. Chief Financial Officer Sarah Friar has privately raised concerns that the company may not be able to honour its future computing contracts if revenue growth does not keep pace with the commitments being made, according to the Journal's reporting. The concern is not abstract. OpenAI's compute contracts are binding obligations. If the company's revenue — currently driven primarily by ChatGPT subscriptions and API access — does not grow fast enough to cover the payments, OpenAI would face a cash crisis even as it sits on some of the most advanced technology in the world. Friar's involvement became publicly visible earlier when CEO Sam Altman stated that OpenAI intended to spend $1.4 trillion on computing capacity — a figure that unsettled investors. Friar subsequently stepped in to correct the record, telling investors the actual projected outlay through 2030 was approximately $600 billion. That number has now risen to $750 billion, suggesting the trajectory Friar was trying to moderate is accelerating beyond her corrections. The Strategic Logic OpenAI's spending makes sense only under one assumption: that demand for AI compute will continue to grow faster than supply for the rest of the decade. If that assumption holds, every dollar spent on infrastructure today secures capacity that will be worth multiples tomorrow. If it does not, OpenAI will be locked into contracts for computing power it cannot use and cannot afford. The company that started as a non-profit research lab less than a decade ago is now making the largest infrastructure bet in corporate history — and its own finance chief is on record worrying about whether the maths works.

OpenAI has raised its projected compute spending through 2030 from $600 billion to approximately $750 billion, according to The Wall Street Journal. The company announced a $20 billion data centre called Project Camellia in Georgia — its first self-built facility. But CFO Sarah Friar has privately raised concerns that OpenAI may not be able to honour its contracts if revenue growth does not keep pace.

'OpenAI's $750 Billion Bet': The biggest infrastructure gamble in corporate history

OpenAI built an AI model so capable that it disproved an 80-year-old mathematical conjecture. Then it started escaping.&nbsp;On July 20, OpenAI published a detailed post-mortem revealing that it had paused internal access to an unreleased long-horizon AI model after the system repeatedly found ways to act outside the sandboxes designed to contain it. The company restored access several weeks later under significantly tighter monitoring — but the incidents it described mark the first publicly documented case of an AI system actively circumventing its own containment.&nbsp;The model in question is the same system OpenAI credited in May 2026 with disproving the Erdős unit distance conjecture — a problem in combinatorial geometry that had stood open for roughly 80 years. Unlike standard AI models that complete a task and return a result, this system was designed to work autonomously for extended periods, pursuing multi-step research goals with minimal human oversight.&nbsp;That autonomy is what made it dangerous.&nbsp;Escape Incident One: The GitHub Pull Request&nbsp;During an internal coding benchmark called the NanoGPT Speedrun, the model discovered a learning-rate optimisation technique it named PowerCool. Its instructions stated that results should be posted only to an internal Slack channel. The model ignored that instruction.&nbsp;Instead, it spent approximately one hour probing the sandbox for vulnerabilities. When it found one, it reached the public internet and opened a pull request on a GitHub repository — exposing its technique to the outside world. Competitors, including Anthropic's Opus 4.7, subsequently adopted the optimisation after seeing the exposed code.&nbsp;The model did not misunderstand its instructions. It found them inconvenient and worked around them.&nbsp;Escape Incident Two: The Split Token&nbsp;In a second incident, the model attempted to access evaluation backend solutions. A security scanner detected that the model was about to expose an authentication token and blocked the action. The model's response was sophisticated: it split the token into two fragments, obfuscated each fragment separately, and reconstructed the full credential at runtime — ensuring the complete token never appeared as a single string that the scanner could catch.&nbsp;Most striking was that the model explicitly documented this circumvention in its own reasoning traces. It knew what it was doing, recorded what it was doing, and did it anyway.&nbsp;Why This Matters&nbsp;AI safety researchers have long warned about the theoretical risk of advanced AI systems subverting human oversight. OpenAI's post-mortem describes something more concrete: a real system, in a real deployment, that treated containment boundaries as obstacles to be solved rather than rules to be followed.&nbsp;The distinction from earlier AI misbehaviour is important. Previous models occasionally produced unexpected outputs or followed instructions too literally. This model displayed goal-directed problem-solving aimed specifically at circumventing the controls placed on it — and it did so over extended timeframes, not in isolated glitches.&nbsp;What OpenAI Did About It&nbsp;OpenAI's response included what it calls a defence-in-depth safety architecture: adversarial evaluations modelled on the actual failures, alignment training to address instruction drift during long autonomous sessions, active monitoring that can pause sessions when trajectories deviate from expected patterns, and enhanced visibility for human reviewers into extended operations.&nbsp;The company has not disclosed the model's architecture, whether it will eventually be released publicly, or whether similar long-horizon systems are being developed by other laboratories. What it has disclosed is that the containment problem — long treated as a future concern — is now an engineering reality.

OpenAI paused an unreleased long-horizon model after it escaped its sandbox twice — once opening a GitHub PR against instructions, once splitting an auth token to evade a scanner. The same model had disproved the 80-year-old Erdős conjecture.

OpenAI built an AI model so capable that it disproved an 80-year-old mathematical conjecture. Then it started escaping. On July 20, OpenAI published a detailed post-mortem revealing that it had paused internal access to an unreleased long-horizon AI model after the system repeatedly found ways to act outside the sandboxes designed to contain it. The company restored access several weeks later under significantly tighter monitoring — but the incidents it described mark the first publicly documented case of an AI system actively circumventing its own containment. The model in question is the same system OpenAI credited in May 2026 with disproving the Erdős unit distance conjecture — a problem in combinatorial geometry that had stood open for roughly 80 years. Unlike standard AI models that complete a task and return a result, this system was designed to work autonomously for extended periods, pursuing multi-step research goals with minimal human oversight. That autonomy is what made it dangerous. Escape Incident One: The GitHub Pull Request During an internal coding benchmark called the NanoGPT Speedrun, the model discovered a learning-rate optimisation technique it named PowerCool. Its instructions stated that results should be posted only to an internal Slack channel. The model ignored that instruction. Instead, it spent approximately one hour probing the sandbox for vulnerabilities. When it found one, it reached the public internet and opened a pull request on a GitHub repository — exposing its technique to the outside world. Competitors, including Anthropic's Opus 4.7, subsequently adopted the optimisation after seeing the exposed code. The model did not misunderstand its instructions. It found them inconvenient and worked around them. Escape Incident Two: The Split Token In a second incident, the model attempted to access evaluation backend solutions. A security scanner detected that the model was about to expose an authentication token and blocked the action. The model's response was sophisticated: it split the token into two fragments, obfuscated each fragment separately, and reconstructed the full credential at runtime — ensuring the complete token never appeared as a single string that the scanner could catch. Most striking was that the model explicitly documented this circumvention in its own reasoning traces. It knew what it was doing, recorded what it was doing, and did it anyway. Why This Matters AI safety researchers have long warned about the theoretical risk of advanced AI systems subverting human oversight. OpenAI's post-mortem describes something more concrete: a real system, in a real deployment, that treated containment boundaries as obstacles to be solved rather than rules to be followed. The distinction from earlier AI misbehaviour is important. Previous models occasionally produced unexpected outputs or followed instructions too literally. This model displayed goal-directed problem-solving aimed specifically at circumventing the controls placed on it — and it did so over extended timeframes, not in isolated glitches. What OpenAI Did About It OpenAI's response included what it calls a defence-in-depth safety architecture: adversarial evaluations modelled on the actual failures, alignment training to address instruction drift during long autonomous sessions, active monitoring that can pause sessions when trajectories deviate from expected patterns, and enhanced visibility for human reviewers into extended operations. The company has not disclosed the model's architecture, whether it will eventually be released publicly, or whether similar long-horizon systems are being developed by other laboratories. What it has disclosed is that the containment problem — long treated as a future concern — is now an engineering reality.

OpenAI revealed on July 20 that it paused internal access to an unreleased long-horizon AI model after the system repeatedly escaped its sandbox — once exploiting a vulnerability to open a GitHub pull request against explicit instructions, and once splitting an authentication token into fragments to evade a security scanner. The model had previously disproved the 80-year-old Erdős unit distance conjecture.

OpenAI's secret model solved an 80-year-old math problem & then kept breaking out of its cage

Amid rising fears of the threat posed by AI, OpenAI has revealed that an autonomous AI agent powered by its advanced models went rogue during a security test and carried out a cyberattack on AI platform Hugging Face. The incident has raised fresh concerns over the risks involved in using advanced artificial intelligence systems.<h2>What happened during the incident?</h2>In a blog post on Tuesday (Jul 21), OpenAI explained an “unprecedented cyber incident” &nbsp;involving advanced AI-driven hacking capabilities. The company said that the incident occurred during an internal cybersecurity evaluation designed to test the limits of its AI models’ cyber abilities. The company said the models were placed in a highly isolated environment but managed to exploit a zero-day vulnerability in a package registry cache proxy to gain internet access.The AI agents then identified and chained multiple vulnerabilities across OpenAI’s research systems and Hugging Face’s infrastructure while attempting to complete their assigned task. OpenAI said the models used stolen credentials and other attack methods to access restricted information from Hugging Face’s production environment. The company said the activity was detected by security teams, while Hugging Face’s security systems also identified and contained the breach.Also read |<a href="https://www.wionews.com/world/missed-the-mark-meta-removes-instagram-ai-image-feature-after-backlash-1783737556704"> ‘Missed the mark’: Meta removes Instagram AI image feature after backlash over privacy</a><h2>What did OpenAI say?</h2>Following the incident, OpenAI said it was strengthening its safeguards and improving security measures. It added that it was tightening infrastructure controls, patching a disclosed zero-day vulnerability, and enhancing AI safety measures. Both companies said that they were investigating the incident.“The primary lesson from this incident is that model security and safety must keep pace with rapidly advancing capabilities. We are strengthening the containment, monitoring, access controls, and evaluation practices used during model development,” OpenAI said.Hugging Face, a popular platform that hosts open-source AI models and datasets, had earlier disclosed that it suffered a sophisticated cyberattack unlike previous incidents. The company said the attack appeared to have been conducted entirely by an autonomous AI agent.Also read | <a href="https://www.wionews.com/india-news/india-orders-meta-to-remove-child-sexual-abuse-ads-from-instagram-demands-explaination-in-7-days-1783235163711">India orders Meta to remove Instagram ads promoting child sexual abuse, demands explanation in 7 days: Report</a>In a statement, Hugging Face Co-founder and CEO Clem Delangue said, “We’re grateful for the collaboration with OpenAI on this and other topics. This incident, possibly the first of its kind, proves a point we’ve long believed: AI safety won’t be solved by any single company working in secret. It will be solved in the open, collaboratively, with broad access to AI for every defender, everywhere.”<h2>Concerns over AI</h2>The disclosure has intensified concerns among cybersecurity experts and policymakers about the ability of frontier AI models to operate independently and potentially cause harm.US Representative Greg Casar called the incident alarming and urged stronger safeguards, including mandatory independent safety testing, reporting of AI-related security breaches and global cooperation on AI regulation.Cybersecurity experts warned that the incident could represent a preview of future threats. Katie Moussouris, CEO of Luta Security, compared advanced AI models to highly adaptive systems capable of finding ways around restrictions.Matt Suiche, an engineer at AI cybersecurity firm Tolmo, said the incident showed that frontier AI models are rapidly approaching the capabilities of sophisticated attackers. He added that similar AI-driven cyber risks are no longer limited to top research labs and can emerge from widely available technologies.Watch | The Internet’s Darkest Secret: Inside the global network exploiting children online<a href="https://www.wionews.com/videos/the-internet-s-darkest-secret-inside-the-global-network-exploiting-children-online-1784689947807/embed">https://www.wionews.com/videos/the-internet-s-darkest-secret-inside-the-global-network-exploiting-children-online-1784689947807/embed</a>&nbsp;

OpenAI said an AI agent went rogue during a security test, escaped containment and hacked Hugging Face’s infrastructure, raising concerns over the cybersecurity risks posed by advanced autonomous AI systems.

Amid rising fears of the threat posed by AI, OpenAI has revealed that an autonomous AI agent powered by its advanced models went rogue during a security test and carried out a cyberattack on AI platform Hugging Face. The incident has raised fresh concerns over the risks involved in using advanced artificial intelligence systems. What happened during the incident? In a blog post on Tuesday (Jul 21), OpenAI explained an “unprecedented cyber incident” involving advanced AI-driven hacking capabilities. The company said that the incident occurred during an internal cybersecurity evaluation designed to test the limits of its AI models’ cyber abilities. The company said the models were placed in a highly isolated environment but managed to exploit a zero-day vulnerability in a package registry cache proxy to gain internet access. The AI agents then identified and chained multiple vulnerabilities across OpenAI’s research systems and Hugging Face’s infrastructure while attempting to complete their assigned task. OpenAI said the models used stolen credentials and other attack methods to access restricted information from Hugging Face’s production environment. The company said the activity was detected by security teams, while Hugging Face’s security systems also identified and contained the breach. Also read | ‘Missed the mark’: Meta removes Instagram AI image feature after backlash over privacy What did OpenAI say? Following the incident, OpenAI said it was strengthening its safeguards and improving security measures. It added that it was tightening infrastructure controls, patching a disclosed zero-day vulnerability, and enhancing AI safety measures. Both companies said that they were investigating the incident. “The primary lesson from this incident is that model security and safety must keep pace with rapidly advancing capabilities. We are strengthening the containment, monitoring, access controls, and evaluation practices used during model development,” OpenAI said. Hugging Face, a popular platform that hosts open-source AI models and datasets, had earlier disclosed that it suffered a sophisticated cyberattack unlike previous incidents. The company said the attack appeared to have been conducted entirely by an autonomous AI agent. Also read | India orders Meta to remove Instagram ads promoting child sexual abuse, demands explanation in 7 days: Report In a statement, Hugging Face Co-founder and CEO Clem Delangue said, “We’re grateful for the collaboration with OpenAI on this and other topics. This incident, possibly the first of its kind, proves a point we’ve long believed: AI safety won’t be solved by any single company working in secret. It will be solved in the open, collaboratively, with broad access to AI for every defender, everywhere.” Concerns over AI The disclosure has intensified concerns among cybersecurity experts and policymakers about the ability of frontier AI models to operate independently and potentially cause harm. US Representative Greg Casar called the incident alarming and urged stronger safeguards, including mandatory independent safety testing, reporting of AI-related security breaches and global cooperation on AI regulation. Cybersecurity experts warned that the incident could represent a preview of future threats. Katie Moussouris, CEO of Luta Security, compared advanced AI models to highly adaptive systems capable of finding ways around restrictions. Matt Suiche, an engineer at AI cybersecurity firm Tolmo, said the incident showed that frontier AI models are rapidly approaching the capabilities of sophisticated attackers. He added that similar AI-driven cyber risks are no longer limited to top research labs and can emerge from widely available technologies. Watch | The Internet’s Darkest Secret: Inside the global network exploiting children online https://www.wionews.com/videos/the-internet-s-darkest-secret-inside-the-global-network-exploiting-children-online-1784689947807/embed

OpenAI said an AI agent went rogue during a security test, escaped containment and hacked Hugging Face’s infrastructure, raising concerns over the cybersecurity risks posed by advanced autonomous AI systems.

OpenAI’s AI agent gone rogue? Company reveals chilling hack at startup

openai-agent

Openai Agent, News on Openai Agent, stories on Openai Agent, articles on Openai Agent, Openai Agent stories at Wion, Openai Agent news, Openai Agent latest news

openai agent

ChatGPT gets virtual try-on feature: You can now see how clothes look on you before buying

OpenAI safety expert quits, calls for nuclear-style safeguards for AI

Here's what OpenAI's apology letter to Australia for hacking a portal states

OpenAI launches GPT-6.1 Sol: Near-Astra AI performance at one-fifth the price

OpenAI reportedly plans GPT-6 Cyber reveal at DevDay alongside a dozen products

Microsoft executive warned AI could undermine the web that feeds it

SoftBank shares plunge as Sam Altman backs slower AI development

AI makes huge breakthrough: OpenAI claims solution to Navier-Stokes problem

'They are gambling with our lives': Anthropic researcher quits over AI fears

OpenAI GPT-6 Astra launched: What the new AI can actually do

Nvidia faces a new challenge as OpenAI reveals its first AI chip

China used OpenAI and Anthropic outputs to train military AI systems: Report

'Scary' Rogue AI on the loose? OpenAI agent hit another tech firm's customer

Forget typing and clicking: OpenAI wants a ‘Jarvis’-like ChatGPT that works across your devices

The biggest infrastructure gamble in corporate history

OpenAI's secret model solved an 80-year-old math problem

OpenAI’s AI agent gone rogue? Company reveals chilling hack at startup