Instagram hack shows danger of outsourcing tech support to AI

This article appeared as part of the Daily Sensemaker newsletter – one story a day to make sense of the world. To receive it in your inbox, featuring content exclusive to the newsletter, sign up for free here.

Instagram says it has resolved an issue that allowed hackers to break into accounts by asking Meta’s AI support bot to change the email address connected with the account. This followed a brief takeover of Barack Obama’s White House profile, which posted pro-Iran content on Sunday before it was recovered. Andy Stone, a spokesperson for Meta, said that claims the vulnerability was used to infiltrate the accounts of world leaders were “false”. But it shows the danger of outsourcing tech support to AI bots that might be tricked into handing over sensitive information to bad actors. Anthropic is currently rolling out its powerful Mythos AI model, despite fears that its ability to identify critical security flaws could be weaponised by hackers.