• Wion
  • /Technology
  • /AI agents now being used by hackers to steal money, CERT-In issues warning

AI agents now being used by hackers to steal money, CERT-In issues warning

AI agents now being used by hackers to steal money, CERT-In issues warning

Your company’s payment approval system could be the weakest link: CERT-In issues warning Photograph: (Freepik)

Story highlights

CERT-In has issued a critical warning about AI-powered cyberattacks and hackers manipulating business payment workflows to divert money. The agency has urged organisations to strengthen independent payment verification, restrict privileged access and monitor suspicious activity.

India's cybersecurity agency, the Indian Computer Emergency Response Team (CERT-In), has issued a critical cybersecurity warning over criminals using AI-powered tools to automate attacks and manipulate business payment systems to divert money. The November 20 advisory warns that the threat is no longer confined to banks and financial institutions. Technology service providers handling government-linked payments, vendor payouts and partner transactions are also at risk.

The warning highlights a worrying shift: hackers may not need to break into a bank or directly attack a payment gateway. They can instead manipulate the business records and approval processes that tell an organisation where its money should go.

How are hackers manipulating business payment systems?

According to CERT-In's Advisory CIAD-2026-0062, attackers have targeted business applications and payment workflows to push through unauthorised transactions using otherwise legitimate systems.

They can alter payment-related business records or approval relationships before using established payment processes to execute transactions. Weak independent verification and reconciliation procedures can allow these changes to go unnoticed.

This creates a serious problem for businesses. A payment may appear to have passed through the usual application and approval checks, even though the underlying beneficiary or payment instructions have been manipulated. CERT-In said incidents have involved organisations outside the banking and financial services sector, including service providers supporting government-linked payment processes and platforms handling vendor or partner payouts. The agency did not disclose the names of affected organisations or the value of the transactions.

Trending Stories

AI agents can automate multiple stages of cyberattacks

The advisory also flags the use of AI-driven agentic offensive frameworks (AOFs) in cyberattacks targeting enterprises, technology service providers, government organisations and academic institutions. These frameworks can automate tasks that attackers would otherwise perform manually. They can scan exposed systems, identify possible entry points and help carry out different stages of an intrusion. CERT-In's threat-hunting activities identified their use across multiple sectors. The concern is that automation can reduce the time and effort needed to identify weaknesses and attempt unauthorised access.

What has CERT-In advised businesses to do?

The agency has recommended stronger checks around payment records, user accounts and third-party service providers.

Key recommendations include:

- Verify payment records independently: Cross-check beneficiary details and payment instructions against trusted records outside the service provider's environment.

- Add independent approval checks: Require separate verification for changes to beneficiaries, payment instructions and approval arrangements.

- Restrict privileged access: Limit administrative accounts, credentials and system privileges, particularly across different client environments.

- Monitor suspicious activity: Track changes to payment records alongside administrator actions, approvals and subsequent transactions.

CERT-In has also reminded organisations that applicable cyber incidents must be reported within six hours of noticing them or being informed about them, in line with its 2022 directions. People facing unauthorised financial transactions can report cyber financial fraud through the National Cyber Crime Reporting Portal or helpline 1930. The warning underlines a broader challenge for businesses: securing a payment system is not enough if the records and approvals behind a transaction can be manipulated. As AI-powered attack tools become more capable, organisations will need to verify not just whether a payment was authorised by a system, but whether the underlying instruction was genuine.

About the Author

Abhinav Yadav

Abhinav is a versatile and adaptive journalist who covers defence, space, and technology for WION. He specialises in breaking down complex subjects into clear, engaging stories tha...Read More