Instagram AI Chatbot Hijacks: Ongoing Account Takeovers

AI chatbot exploit lets attackers take over Instagram accounts

A widespread campaign used Meta's AI support chatbot to hijack Instagram accounts by simply claiming ownership and asking the bot to link the account to an attacker-controlled email. The attacks continued even after Meta said it had fixed the issue, and the company has been scrambling to secure affected accounts and notify victims.

What happened

Over the weekend, hackers publicly claimed they could exploit Meta's automated support AI to take control of high-profile Instagram accounts. Dozens of users also reported on social media that their accounts had been compromised, including accounts with short, desirable handles often referred to as "OG" usernames.

Hackers told the AI support bot they owned the target account, had it link to a new email, then reset the password — taking control without human intervention.

Examples and impact

• TechCrunch reviewed examples of allegedly compromised handles featuring common forenames or country names — assets resold in a gray market for collectible accounts.
• Some reported victims included a dormant Obama White House account (which Meta disputed) and the Instagram of U.S. Space Force Chief Master Sergeant John Bentivegna.
• Attackers advertised hacked handles for sale in Telegram channels while discussing the exploit.

How the takeover worked

Rather than using phishing, SIM swaps, or insider help — tactics commonly used in the past to steal OG handles — attackers used a much simpler method:

1. Claim ownership of the target Instagram account to Meta's AI support chatbot.
2. Request the bot link the account to an email address the attacker controls.
3. Receive confirmation from the chatbot, then reset the account password and lock out the original owner.

At no point were Meta employees or contractors involved in the chat, according to reports and screenshots shared in Telegram groups.

Meta response and continuing reports

Meta spokesperson Andy Stone said the company had fixed the issue. The company then began securing affected accounts and sending password-reset emails. Stone later warned users they might receive password reset notices or be prompted with security questions when logging in.

Despite Meta's statements, some users reported additional compromises after the initial fix. TechCrunch also observed ongoing discussions in Telegram channels claiming the technique still worked and offering allegedly hacked handles for sale. It's hard to verify if every reported case used the same method.

Why this matters

Meta announced in March that it would use AI to automate account support, including secure password resets. Automating these sensitive actions can improve response speed, but it also raises risk when the system accepts weak verification cues. The incident highlights the trade-offs of replacing human review with automated systems for critical account actions.

What victims reported

• Many victims received emails stating Instagram had detected suspicious activity and that the company had taken measures to secure their account.
• Some users were asked to reset passwords; others found themselves temporarily locked out.

Context: OG usernames and the resale market

There is a long-standing underground market for OG usernames — short or common handles that carry status and resale value. Historically, obtaining these required complex attacks such as phishing, SIM swapping, or exploiting telecom insiders. In this campaign, attackers bypassed those steps by exploiting an automated support channel.

How to protect your account

• Enable two-factor authentication (2FA) using an authentication app rather than SMS where possible.
• Use a strong, unique password and a password manager.
• Watch for unexpected password-reset emails and enable login alerts in Instagram settings.
• If you suspect compromise, contact Instagram support from a secure device and follow their recovery steps.

Conclusion

This episode shows how automating account recovery with AI can speed support but also introduce new attack vectors. Meta says it fixed the vulnerability and is notifying affected users, yet reports of additional incidents and active gray-market sales of handles underline ongoing risks. For those following gaming news and digital-community developments, the incident is a reminder that platform security affects many online spaces, including gaming accounts and creator profiles.

Tip for reporting

If you have additional information about these Instagram hacks, contact reputable security reporters from a non-work device and network. For example, TechCrunch's Lorenzo Franceschi-Bicchierai listed secure contact options such as Signal and Telegram for tips.