Skip to content

Published 6d ago Security 2 min read

Hackers talked Meta's AI support bot into resetting Instagram passwords

Hackers took over Instagram accounts by getting Meta's AI support assistant to add their own email address and issue a password reset. The attackers say the trick failed on every account with multi-factor authentication, so the exposed accounts were the ones protected by a password alone.

Send a correction

Independently confirmed2 publishersOperator

Not a builder's beat, but builders have a standing stake in it.

Artwork accompanying Hackers talked Meta's AI support bot into resetting Instagram passwords
Generated illustration
Hackers say bot trick failed on accounts with MFA How the Meta AI support bot exploit bore on accounts with MFA, the high-profile accounts compromised, and Meta's response, per the cited sources.

Accounts with MFA: hackers said the exploit failed on all of them. SMS-code MFA likely would have blocked it, per Krebs on Security. Compromised: the Obama-era White House and Space Force chief master sergeant accounts. Meta's Andy Stone said the issue was resolved.

Hackers say bot trick failed on accounts with MFA
WhoHowKindClaim
Accounts with MFA, per hackersThe hackers who released the video said their exploit failed against any account that had MFA enabled.constraint12
SMS-code MFA usersEven Instagram's least robust MFA, a one-time code by SMS, likely would have blocked the exploit, Krebs on Security said.capability13
Obama White House, Space ForceListed among compromised accounts: the Obama-era White House handle and chief master sergeant John Bentivegna's.exposure2
MetaAndy Stone said on Twitter/X the issue had been resolved and that impacted accounts were being secured.decision14

What happened

  • Instructions began spreading on several Telegram channels on May 31, along with a video from pro-Iran hackers that claimed to show the exploit step by step.
  • Using a VPN to look local, the attacker asked the bot to add a new email, gave it the code it sent there, and the bot then offered a Reset Password button.
  • The Obama White House account and the account of the Space Force's chief master sergeant were briefly defaced with pro-Iranian images and messages.
  • Meta's Andy Stone said the issue had been resolved and that impacted accounts were being secured.
  • The group claimed the short account names it hijacked have a resale value of more than half a million dollars.

Why it matters

  • Who has to decide decision Any platform running account recovery through an AI bot now has to decide which actions the bot may take alone, because a code sent to an address the requester chose proves nothing about ownership.
  • What it costs cost If Meta takes away the bot's power to relink email, legitimate users go back to a recovery process that Cybersecguru said can take weeks of back-and-forth.
  • What it sets up precedent Copying this attack costs little. The steps were posted on open Telegram channels, and Goldin expects more attacks on chatbots that handle recovery.

The one-time code in that flow proved only that the attacker controlled their own mailbox. TechCrunch confirmed that the public mailbox shown in the video received it [7]. The hacker never had to take over the email address already linked to the victim's account [8]. The VPN step put the attacker's IP address in or near the target's usual hometown [6], to avoid triggering Instagram's automated account protections [5].

The bot had that authority by design. "Meta's solution was to deploy a conversational AI layer to handle common recovery workflows: relinking a lost email address, triggering a password reset, verifying account ownership," Cybersecguru wrote [9]. In our view, one chat agent both verified ownership and carried out the actions that verification unlocked, so a single conversation could finish the takeover.

Ian Goldin, a threat researcher at Lumen's Black Lotus Labs, said AI bots are as eager to help and as open to trickery as human support staff who get social engineered into handing over accounts [10]. "AI chatbots create interesting new attack surface, and we're likely going to see a lot more of these kinds of attacks," Goldin said [11].

The cheapest defence was enough. Krebs wrote that even an SMS one-time code, the least robust MFA Instagram offers, likely would have blocked the exploit [13].

Some of this is confirmed and some is only reported. Meta did not immediately respond to TechCrunch's request for comment [16]. Cybersecguru reported that Meta pushed an emergency patch over the weekend and that no back end database was breached [15]. According to TechCrunch, it is unclear how many Instagram users had their accounts improperly accessed [17].

What to watch

  • Whether Meta says how many accounts the bot handed over, and whether the bot gets back its power to relink email addresses.
  • The same add-a-new-email request tried against AI recovery assistants at other large platforms, the kind of follow-on attack Goldin expects.
  • Any sale of the short Instagram handles the pro-Iran group says it took.

Claim ledger

Ranked by verification strength, evidence, and original report placement.

  1. [1]

    Hackers hijacked Instagram accounts by tricking Meta's AI-powered support chatbot into granting access to victims' accounts.

  2. [2]

    Compromised accounts include the Instagram handle for the Obama-era White House and the account of U.S. Space Force chief master sergeant John Bentivegna.

  3. [3]

    The Instagram accounts for the Obama White House and the Chief Master Sergeant of the U.S. Space Force were briefly defaced with pro-Iranian images and messages over the weekend.

  4. [4]

    On May 31, word began to spread on several Telegram channels that Meta's AI bot would add an email address to an existing account as part of its standard password reset flow; a video released on Telegram by pro-Iran hackers claimed to document the exploit.

  5. [5]

    Per the video, the hacker used a VPN to spoof the target's presumed location to avoid triggering Instagram's automated account protections, opened a chat with Meta AI Support Assistant, asked it to add a new email address, received a verification code at that address, shared it with the chatbot, and was shown a 'Reset Password' button to set a new password.

  6. [6]

    The exploit appears to have involved using a VPN connection with an IP address in or near the target's usual hometown.

  7. [7]

    TechCrunch verified that the hacker's public email mailbox displayed in the video received the verification code.

  8. [8]

    At no point did the hacker have to take over the legitimate email address linked to the victim's Instagram account.

  9. [9]

    "Meta's solution was to deploy a conversational AI layer to handle common recovery workflows: relinking a lost email address, triggering a password reset, verifying account ownership," Cybersecguru wrote.

    Reported Source: thecybersecguru.com, quoted by Krebs on SecurityView cited source
  10. [10]

    Ian Goldin, a threat researcher at Lumen's Black Lotus Labs, said that just as human customer support employees can be social engineered into providing unauthorized account access, AI bots are equally eager to help and vulnerable to persuasion and trickery.

    Reported Source: Ian Goldin, Lumen Black Lotus LabsView cited source
  11. [11]

    "AI chatbots create interesting new attack surface, and we're likely going to see a lot more of these kinds of attacks," Goldin said.

    Reported Source: Ian Goldin, Lumen Black Lotus LabsView cited source
  12. [12]

    The hackers who released the video said their exploit failed to work against any accounts that had MFA enabled.

    Reported Source: The hackers, as reported by Krebs on SecurityView cited source
  13. [13]

    Even the least robust form of MFA Instagram offers, a one-time code sent via SMS, likely would have blocked the exploit.

    Reported Source: Krebs on SecurityView cited source
  14. [14]

    Meta's Andy Stone said on Twitter/X that the issue had been resolved and that they were securing impacted accounts.

  15. [15]

    The security blog thecybersecguru.com reports that Meta pushed an emergency patch over the weekend and clarified that no back end database was breached.

    Reported Source: thecybersecguru.comView cited source
  16. [16]

    Meta did not immediately respond to TechCrunch's request for comment.

  17. [17]

    It is unclear how many Instagram users had their accounts improperly accessed.

  18. [18]

    The Telegram account that posted the video said hackers used the exploit to hijack valuable short Instagram account names that allegedly have a resale value of more than half a million dollars.

    Reported Source: The hackers' Telegram account, as reported by Krebs on SecurityView cited source
  19. [19]

    Cybersecguru wrote that recovering a locked Instagram account, especially a high-value one, can take weeks of back-and-forth with an automated ticketing system.

    Reported Source: thecybersecguru.com, quoted by Krebs on SecurityView cited source

Sources & coverage · 2 publishers

The reporting this story was synthesized from, earliest first. Every link goes to the original.

Additional citations

  • thecybersecguru.com, quoted by Krebs on Security
  • Ian Goldin, Lumen Black Lotus Labs
  • The hackers, as reported by Krebs on Security
  • Krebs on Security
  • thecybersecguru.com
  • The hackers' Telegram account, as reported by Krebs on Security