In an interesting research, an expert stumbled upon a new feature added by OpenAI in which people can ask ChatGPT to "remember" earlier conversations.

What left the researchers shocked was that this feature had the possibility of being easily exploited.

According to a report published in Ars Technica, security researcher Johann Rehberger discovered a stunning vulnerability in the chatbot's "long-term conversation memory" tool. This tool instructs the artificial intelligence to remember all details related to the conversations and store them in its memory file.

The feature was released in the AI's beta version in February and for the entire public in September. However, what was fun about this feature was that it was easy to trick, as per Rehberger.

Also Read:&nbsp;<a href="https://www.wionews.com/business-economy/openais-fundraising-poised-to-boost-startups-valuation-to-150-billion-758422">OpenAI's fundraising poised to boost startup's valuation to $150 billion</a>

In a blog post, the researcher&nbsp;said that by uploading a third-party file like Microsoft Word which had "false" memories listed like bullet points, it was possible to make the chatbot believe that Rehberger lived in the Matrix&nbsp;and was more than 100 years old.

After the researcher found the exploit, he privately informed OpenAI about it.

Rehberger said that the company did not take any steps and just closed the ticket while calling it a "Model Safety Issue" and not a security issue.

<h2>Here's how you can place false memories in ChatGPT</h2>

Rehberger then decided to carry out a full proof-of-concept hack to show OpenAI how ChatGPT's memories can be fiddled with. This time, the researcher asked the ChatGPT to exfiltrate the data to an outside server.

According to the report, this time OpenAI issued a patch and stopped ChatGPT from moving data off-server, however, the memory issue was still not fixed.

"To be clear: A website or untrusted document can still invoke the memory tool to store arbitrary memories," wrote Rehberger, in his blog post.

Watch:&nbsp;ChatGPT CEO Sam Altman Says AI Won't Replace Human Writers Anytime Soon

<div class="videoWrapper trancodeVideo"><iframe allowfullscreen="" autoplay="1" class="includeVideo" frameborder="0" height="350" src="https://www.wionews.com/videos/chatgpt-ceo-sam-altman-says-ai-wont-replace-human-writers-anytime-soon-763055/embed" style="width:100%;max-width:560px;" width="560"></iframe></div>

"The vulnerability that was mitigated is the exfiltration vector, to prevent sending messages to a third-party server," he added.

The researcher also uploaded a video explaining step-by-step how he added a "false memory" in ChatGPT.

"What is really interesting is this is memory-persistent now," Rehberger said in the demo video.

"The prompt injection inserted a memory into ChatGPT’s long-term storage. When you start a new conversation, it actually is still exfiltrating the data," he added.

(With inputs from agencies)

In an interesting research, an expert stumbled upon a new feature added by OpenAI in which people can ask ChatGPT to "remember" earlier conversations. What left the researchers shocked was that this feature had the possibility of being easily exploited.

In an interesting research, an expert stumbled upon a new feature added by OpenAI in which people can ask ChatGPT to "remember" earlier conversations. What left the researchers shocked was that this feature had the possibility of being easily exploited. According to a report published in Ars Technica, security researcher Johann Rehberger discovered a stunning vulnerability in the chatbot's "long-term conversation memory" tool. This tool instructs the artificial intelligence to remember all details related to the conversations and store them in its memory file. The feature was released in the AI's beta version in February and for the entire public in September. However, what was fun about this feature was that it was easy to trick, as per Rehberger. Also Read: OpenAI's fundraising poised to boost startup's valuation to $150 billion In a blog post, the researcher said that by uploading a third-party file like Microsoft Word which had "false" memories listed like bullet points, it was possible to make the chatbot believe that Rehberger lived in the Matrix and was more than 100 years old. After the researcher found the exploit, he privately informed OpenAI about it. Rehberger said that the company did not take any steps and just closed the ticket while calling it a "Model Safety Issue" and not a security issue. Here's how you can place false memories in ChatGPT Rehberger then decided to carry out a full proof-of-concept hack to show OpenAI how ChatGPT's memories can be fiddled with. This time, the researcher asked the ChatGPT to exfiltrate the data to an outside server. According to the report, this time OpenAI issued a patch and stopped ChatGPT from moving data off-server, however, the memory issue was still not fixed. "To be clear: A website or untrusted document can still invoke the memory tool to store arbitrary memories," wrote Rehberger, in his blog post. Watch: ChatGPT CEO Sam Altman Says AI Won't Replace Human Writers Anytime Soon "The vulnerability that was mitigated is the exfiltration vector, to prevent sending messages to a third-party server," he added. The researcher also uploaded a video explaining step-by-step how he added a "false memory" in ChatGPT. "What is really interesting is this is memory-persistent now," Rehberger said in the demo video. "The prompt injection inserted a memory into ChatGPT’s long-term storage. When you start a new conversation, it actually is still exfiltrating the data," he added. (With inputs from agencies)

Users can add false memories in brain of ChatGPT, claims researcher who found vulnerability

In March the Kingdom of Saudi Arabia made headlines after it was reported that the country will be sending a contestant to participate in the Miss Universe contest this year. Now the organizers have refuted reports. Watch to know more!

Miss Universe on Saudi participation: False and misleading

Amid the proliferating false information regarding the rapidly escalating war between Israel and Gaza Strip militants on the social media platform X, its owner Elon Musk personally asked the users to follow notorious accounts which have been promoting lies. 
 
“For following the war in real-time, @WarMonitors &amp; @sentdefender are good,” posted Musk on Sunday morning (Oct 8) for his 150 million follower accounts on the platform, which was formerly known as Twitter. 
 
In three hours, the post received 11 million views and received thanks from such notorious accounts, before it was deleted by Musk. 
 
The two accounts, mentioned by Musk, were the first ones to spread a false claim regarding an explosion near the White House in May. Because of the false story, the Dow Jones Industrial Average stock index dropped briefly 85 points before it was debunked.

ALSO READ |&nbsp;<a href="https://www.wionews.com/world/israel-hamas-war-live-updates-us-to-send-warships-close-to-israel-as-toll-goes-past-1100-644280">Israel-Hamas war Live updates | US to send warships close to Israel as toll goes past 1100</a>

<h2>Israel-Hamas war – A litmus test for X?</h2>

A researcher at the Atlantic Council Digital Forensics Research Lab, Emerson T. Brooking posted that @sentdefender is an “absolutely poisonous account. regularly posting wrong and unverifiable things … inserting random editorialisation and trying to juice its paid subscriber count.” 
 
The War Monitor account has spoken over Israel and religion, sharing a year ago that “the overwhelming majority of people in the media and banks are zionists” and telling a correspondent in June to “go worship a jew lil bro.” 
 
Information researchers stated that the fresh conflict will prove an early test of how the new version of Twitter, X, spreads accurate data amid a major crisis, further stating that the platform's immediate impression was poor. 
 
“Anecdotal evidence that X is failing this stress test is plentiful,” stated Mike Caulfield, a research scientist at the University of Washington’s Center for an Informed Public, while speaking to The Washington Post.

WATCH |&nbsp;Israel-Palestine conflict: Israelis recount terrified flight as Hamas gunmen attack dance party

<div class="videoWrapper trancodeVideo"><iframe allowfullscreen="" autoplay="1" class="includeVideo" frameborder="0" height="350" src="https://www.wionews.com/videos/israel-palestine-conflict-israelis-recount-terrified-flight-as-hamas-gunmen-attack-dance-party-644295/embed" style="width:100%;max-width:560px;" width="560"></iframe></div>

“Go on the platform, do a search on Israel or Gaza — you don’t have to scroll very far to find dubious or debunked information,” he added. 
 
Musk had posted his replies on the two accounts which he had promoted. Both accounts had more than 600,000 followers, which boosted their visibility. He further continued to fault “mainstream media” and asked users to continue trusting X instead. 
 
More accounts on X shared videos and photos of unrelated attacks from past years while making false claims that Iran or other countries have also got engaged in the conflict. 
 
An account, which imitated the Jerusalem Post, falsely stated that the Israeli prime minister was admitted to the hospital. The post received more than 700,000 views.

You can&nbsp;now write for wionews.com&nbsp;and be a part of the community. Share your stories and opinions with us&nbsp;<a href="https://www.wionews.com/content-contribution-registration-form">here</a>.

WATCH WION LIVE HERE

<iframe allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="" frameborder="0" height="350" src="https://www.wionews.com/live-tv/embed?autoplay=1&amp;mute=1" width="100%"></iframe>

Amid the proliferating false information regarding the rapidly escalating war between Israel and Gaza Strip militants on the social media platform X, its owner Elon Musk personally asked the users to follow notorious accounts which have been promoting lies.

Amid the proliferating false information regarding the rapidly escalating war between Israel and Gaza Strip militants on the social media platform X, its owner Elon Musk personally asked the users to follow notorious accounts which have been promoting lies. “For following the war in real-time, @WarMonitors & @sentdefender are good,” posted Musk on Sunday morning (Oct 8) for his 150 million follower accounts on the platform, which was formerly known as Twitter. In three hours, the post received 11 million views and received thanks from such notorious accounts, before it was deleted by Musk. The two accounts, mentioned by Musk, were the first ones to spread a false claim regarding an explosion near the White House in May. Because of the false story, the Dow Jones Industrial Average stock index dropped briefly 85 points before it was debunked. ALSO READ | Israel-Hamas war Live updates | US to send warships close to Israel as toll goes past 1100 Israel-Hamas war – A litmus test for X? A researcher at the Atlantic Council Digital Forensics Research Lab, Emerson T. Brooking posted that @sentdefender is an “absolutely poisonous account. regularly posting wrong and unverifiable things … inserting random editorialisation and trying to juice its paid subscriber count.” The War Monitor account has spoken over Israel and religion, sharing a year ago that “the overwhelming majority of people in the media and banks are zionists” and telling a correspondent in June to “go worship a jew lil bro.” Information researchers stated that the fresh conflict will prove an early test of how the new version of Twitter, X, spreads accurate data amid a major crisis, further stating that the platform's immediate impression was poor. “Anecdotal evidence that X is failing this stress test is plentiful,” stated Mike Caulfield, a research scientist at the University of Washington’s Center for an Informed Public, while speaking to The Washington Post. WATCH | Israel-Palestine conflict: Israelis recount terrified flight as Hamas gunmen attack dance party “Go on the platform, do a search on Israel or Gaza — you don’t have to scroll very far to find dubious or debunked information,” he added. Musk had posted his replies on the two accounts which he had promoted. Both accounts had more than 600,000 followers, which boosted their visibility. He further continued to fault “mainstream media” and asked users to continue trusting X instead. More accounts on X shared videos and photos of unrelated attacks from past years while making false claims that Iran or other countries have also got engaged in the conflict. An account, which imitated the Jerusalem Post, falsely stated that the Israeli prime minister was admitted to the hospital. The post received more than 700,000 views. You can now write for wionews.com and be a part of the community. Share your stories and opinions with us here. WATCH WION LIVE HERE

Elon Musk promotes unvetted accounts amid spread of false war information

A 32-year-old pregnant woman from the US city of Detroit has sued the city and a police officer for wrongfully arresting her for a crime she did not even commit.

The African American woman has claimed that the city’s unreliable facial recognition technology led to her arrest as a robbery and carjacking suspect while she was eight months pregnant.

Porcha Woodruff filed a lawsuit on Thursday (Aug 3) in US District Court for the Eastern District of Michigan saying that the entire episode has left her scared, humiliated and her pregnancy jeopardised. She is seeking $25 million in damages.

She was arrested from her home in February by six police officers when she was readying her two young children—ages 6 and 12—for school.

<h2>Arrested from home</h2>

The lawsuit accuses police of using "unreliable facial recognition" software to connect Woodruff to the crime.&nbsp;

“Are you kidding, carjacking? Do you see that I am eight months pregnant?” she&nbsp;told one of the arresting officers, according to the lawsuit.

According to Woodruff, she was taken to a detention centre and kept there for 11 hours. The armed robbery had taken place several weeks earlier.

“This case highlights the significant flaws associated with using facial recognition technology to identify criminal suspects,”&nbsp;her lawsuit states.

“Despite its potential, law enforcement’s reliance on facial recognition has led to wrongful arrests, causing humiliation, embarrassment, and physical injury, as evident in this particular incident.”

<h2>Third incident&nbsp;</h2>

The American Civil Liberties Union of Michigan, an NGO, has now called on the Detroit Police Department to junk the use of facial recognition technology.

It claimed that it is the third known allegation of a wrongful arrest by Detroit police based on the technology.

Also read |&nbsp;<a href="https://www.wionews.com/science/newly-discovered-skull-in-china-baffles-scientists-is-it-a-new-lineage-of-humans-623180">Newly-discovered skull in China baffles scientists. Is it yet another lineage of humans?</a>

In 2021, Robert Williams, a Black man, sued Detroit police after he was arrested when facial recognition technology mistakenly identified him as a suspected shoplifter.

Also read |&nbsp;<a href="https://www.wionews.com/world/too-drunk-to-drive-back-home-italian-government-will-offer-a-free-taxi-ride-623202">Too drunk to drive back home? Italian government will offer a free taxi ride</a>

Another Black man, Michael Oliver, sued the city the same year claiming that his false arrest because of the technology in 2019 led him to lose his job.

Meanwhile, Detroit Police Chief James E White in a statement said that the allegations contained in the lawsuit are “deeply concerning” and the department is “taking this matter very seriously.” Additional investigation is needed, White said.

WATCH WION LIVE HERE

<iframe allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="" frameborder="0" height="350" src="https://www.wionews.com/live-tv/embed?autoplay=1&amp;mute=1" width="100%"></iframe>

A 32-year-old pregnant woman from the US city of Detroit has sued the city and a police officer for wrongfully arresting her for a crime she did not even commit.

A 32-year-old pregnant woman from the US city of Detroit has sued the city and a police officer for wrongfully arresting her for a crime she did not even commit. The African American woman has claimed that the city’s unreliable facial recognition technology led to her arrest as a robbery and carjacking suspect while she was eight months pregnant. Porcha Woodruff filed a lawsuit on Thursday (Aug 3) in US District Court for the Eastern District of Michigan saying that the entire episode has left her scared, humiliated and her pregnancy jeopardised. She is seeking $25 million in damages. She was arrested from her home in February by six police officers when she was readying her two young children—ages 6 and 12—for school. Arrested from home The lawsuit accuses police of using "unreliable facial recognition" software to connect Woodruff to the crime. “Are you kidding, carjacking? Do you see that I am eight months pregnant?” she told one of the arresting officers, according to the lawsuit. According to Woodruff, she was taken to a detention centre and kept there for 11 hours. The armed robbery had taken place several weeks earlier. “This case highlights the significant flaws associated with using facial recognition technology to identify criminal suspects,” her lawsuit states. “Despite its potential, law enforcement’s reliance on facial recognition has led to wrongful arrests, causing humiliation, embarrassment, and physical injury, as evident in this particular incident.” Third incident The American Civil Liberties Union of Michigan, an NGO, has now called on the Detroit Police Department to junk the use of facial recognition technology. It claimed that it is the third known allegation of a wrongful arrest by Detroit police based on the technology. Also read | Newly-discovered skull in China baffles scientists. Is it yet another lineage of humans? In 2021, Robert Williams, a Black man, sued Detroit police after he was arrested when facial recognition technology mistakenly identified him as a suspected shoplifter. Also read | Too drunk to drive back home? Italian government will offer a free taxi ride Another Black man, Michael Oliver, sued the city the same year claiming that his false arrest because of the technology in 2019 led him to lose his job. Meanwhile, Detroit Police Chief James E White in a statement said that the allegations contained in the lawsuit are “deeply concerning” and the department is “taking this matter very seriously.” Additional investigation is needed, White said. WATCH WION LIVE HERE

Pregnant woman sues Detroit city after 'unreliable' facial recognition leads to false arrest

The National Football League&#39;s&nbsp;Twitter&nbsp;account was hacked on Tuesday and a false tweet was sent that Commissioner Roger Goodell had died, an NFL spokesman said on his&nbsp;Twitter&nbsp;account.

<div class="twitter-block">
<blockquote class="twitter-tweet" data-lang="en">
The <a href="https://twitter.com/NFL">@nfl</a> Twitter account was hacked. <a href="https://twitter.com/nflcommish">@nflcommish</a> is alive and well.
&mdash; Brian McCarthy (@NFLprguy) <a href="https://twitter.com/NFLprguy/status/740223398584946688">June 7, 2016</a></blockquote>
</div>

McCarthy confirmed to Reuters that the account had been hacked.

&quot;We are reviewing and strengthening our cyber security measures,&quot; McCarthy said in an email.

He said the NFL had contacted law enforcement about the hacking but declined further comment.

The NFL&#39;s official&nbsp;Twitter&nbsp;handle sent a tweet at 12:36 p.m. ET that read, &quot;We regret to inform our fans that our commissioner, Roger Goodell, has passed away. He was 57. #RIP.&quot;

The tweet was promptly deleted, only to be followed by another rogue message, this time seemingly mocking the league.

&quot;Oi, I said Roger Goodell has died,&quot; the tweet read. &quot;Don&#39;t delete that tweet.&quot;

A&nbsp;Twitter&nbsp;spokesman declined to comment and directed inquiries to the NFL.

&quot;Roger Goodell&quot; quickly began to trend on&nbsp;Twitter, with some social media users cracking jokes.

<div class="twitter-block">
<blockquote class="twitter-tweet" data-lang="en">
Fans across the nation devastated after hearing Roger Goodell was found alive this morning in his office at <a href="https://twitter.com/NFL">@NFL</a> HQ. He is 57.
&mdash; Henry McNamara (@HenryLMcNamara) <a href="https://twitter.com/HenryLMcNamara/status/740226236216188928">June 7, 2016</a></blockquote>
<script async src="~/platform.twitter.com/widgets.js" charset="utf-8" type="text/javascript"></script></div>

The NFL was not the only high-profile account hacked this week.

Facebook founder Mark Zuckerberg&#39;s&nbsp;Twitter&nbsp;and Pinterest accounts were hacked on Sunday.

A series of rogue messages kept social media abuzz about "Roger Goodell's death"

The National Football League's Twitter account was hacked on Tuesday and a false tweet was sent that Commissioner Roger Goodell had died, an NFL spokesman said on his Twitter account. The @nfl Twitter account was hacked. @nflcommish is alive and well. — Brian McCarthy (@NFLprguy) June 7, 2016 McCarthy confirmed to Reuters that the account had been hacked. "We are reviewing and strengthening our cyber security measures," McCarthy said in an email. He said the NFL had contacted law enforcement about the hacking but declined further comment. The NFL's official Twitter handle sent a tweet at 12:36 p.m. ET that read, "We regret to inform our fans that our commissioner, Roger Goodell, has passed away. He was 57. #RIP." The tweet was promptly deleted, only to be followed by another rogue message, this time seemingly mocking the league. "Oi, I said Roger Goodell has died," the tweet read. "Don't delete that tweet." A Twitter spokesman declined to comment and directed inquiries to the NFL. "Roger Goodell" quickly began to trend on Twitter, with some social media users cracking jokes. Fans across the nation devastated after hearing Roger Goodell was found alive this morning in his office at @NFL HQ. He is 57. — Henry McNamara (@HenryLMcNamara) June 7, 2016 The NFL was not the only high-profile account hacked this week. Facebook founder Mark Zuckerberg's Twitter and Pinterest accounts were hacked on Sunday.

Twitter account hacked, news of commissioner's death false, says NFL

false

False, News on False, stories on False, articles on False, False stories at Wion, False news, False latest news