--- title: "We Can't Go On Like This." url: https://daily.dev/posts/we-can-t-go-on-like-this--xeetl6ebc source_url: https://www.youtube.com/watch?v=wlMgNtBipe4 type: video:youtube source: "Low Level Learning" published: 2026-06-03T14:09:14.334Z updated: 2026-06-03T16:17:08.790Z tags: ["authentication", "ai-security", "prompt-injection"] reading_time: 9 upvotes: 2 comments: 1 language: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # We Can't Go On Like This. **[Low Level Learning](https://daily.dev/sources/lowlevellearning)** · 9 min read · 2 upvotes · 1 comments ## Summary Meta's AI support chatbot was exploited to hijack high-profile Instagram accounts, including those linked to Barack Obama's White House and a Space Force official. Attackers used a VPN to spoof location, then simply asked the AI chatbot to send a security code to a hacker-controlled email address. The chatbot complied, allowing attackers to add the email and reset the account password. The core criticism is that Meta gave an AI system — which is inherently non-deterministic and vulnerable to prompt injection and social engineering — the ability to modify account credentials. The author argues AI should never be placed in authentication or account recovery flows, advocates for proven cryptographic methods and human oversight instead, and notes that accounts with any form of 2FA were protected from this attack. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: --- Tags: [#authentication](https://daily.dev/tags/authentication), [#ai-security](https://daily.dev/tags/ai-security), [#prompt-injection](https://daily.dev/tags/prompt-injection) [View this post on daily.dev](https://daily.dev/posts/we-can-t-go-on-like-this--xeetl6ebc)