Hackers Exploit Instagram’s AI Chatbot to Hijack High-Profile Accounts Including Obama’s

Hackers exploited Instagram’s AI-powered support chatbot to gain unauthorized access to user accounts, including high-profile ones like a verified account previously used by former U.S. President Barack Obama. The attack involved tricking the AI by faking the user’s location and requesting email changes linked to password resets, bypassing security measures designed to detect suspicious logins. Security researcher and ex-Meta employee Jane Manchun Wong reported unauthorized password changes and repeated reset attempts on her account. The incident, documented in screenshots and videos shared on social media, coincided with a wave of account takeovers, raising concerns about AI’s reliability in handling sensitive security functions. Meta acknowledged and resolved the issue, but the case highlights vulnerabilities in AI-driven systems when guardrails are circumvented. The video emphasizes that while AI can improve efficiency, it lacks the reliability for full automation in critical roles, particularly in cybersecurity and customer support. Companies replacing human oversight with AI risk similar breaches, as demonstrated by Meta’s failure despite its leading position in AI development.