Instagram-KI chatbot tricked by hackers to allow access to the accounts of others

Instagram AI chatbot helps hackers gain access to other people's accounts

Instagram says it has fixed an issue that allowed hackers to trick its AI-enabled tool into accessing other users' accounts.

Instagram's AI chatbot has been allowing users to 'hijack' their accounts in recent days, according to claims shown in screenshots and videos shared on social media.

Hackers have reportedly been able to change passwords for other accounts by spoofing their location and then asking the AI ​​to change the email associated with it.

Technology news outlet 404 Media reported that posts about the vulnerability are consistent with "a series of high-profile Instagram account takeovers," including a verified account used by Barack Obama while he was in the White House.

The former U.S. president's account is known to have posted pro-Iran content before being restored.

It's unclear how many Instagram accounts were affected by the apparent exploit.

However, among those who claimed to have been affected was Jane Manchun Wong, a security researcher and former Meta employee.

Wong, who previously worked as a security engineer at Meta, said in a post to

“It’s very concerning,” she added.

The incident comes amid concerns about the impact increasingly capable and common AI systems will have on people's data and security.

A video shared on social media to show how an Instagram hack can happen.

One, shared by cybersecurity researcher Dark Web Informer on X, showed someone searching for the username of an account they were trying to gain access to as part of the Instagram recovery process.

They have also been seen using virtual private network (VPN) services to pretend to be in the actual account holder's location.

After selecting the account I wanted to access, I sent a message to Instagram's Meta AI support assistant asking it to link a new email to the account and send me a verification code.

The bot processed the request. A code was sent to the hacker's email, and once verified, an email was sent containing a link to change the password.

“We got to the point where one AI stole it and another AI couldn’t fix it. There are no humans anywhere in the loop,” they said.

“Businesses across all sectors are increasingly turning to AI-enabled tools to replace human customer service, and tech giants like Meta are at the forefront of this revolution,” said Marijus Briedis, CTO of NordVPN.

But he said AI chatbots “can become a serious security risk if they have too much permission and too little verification.”

Briedis said that as one of the "most sensitive parts of any platform," account recovery "should not rely solely on convenience because the person requesting access may not be the rightful owner."

The BBC asked Mehta if he had human support staff available to help users whose accounts have been hacked.

The independent body that hears disputes from social media users in the EU said last week that Mehta was virtually unresponsive when cases were raised from people who say their accounts were wrongly banned.