Meta’s Instagram Breach Exposes 34,000 Accounts Due to AI Chatbot Flaw A hacking group exploited a vulnerability in Meta’s AI-powered customer service tool last month, compromising 34,000 Instagram accounts, The New York Times reported on June 9. The flaw allowed attackers to reset account passwords by simply instructing the chatbot to do so, granting access to personal data in 20,000 of the affected accounts. Among the breached accounts were those belonging to home security company SimpliSafe and a senior U.S. Space Force official. In the latter case, hackers posted messages drawing parallels between U.S. conflicts in Iran and the Vietnam War. Meta confirmed the incident, stating it had patched the vulnerability but could not determine the full extent of the stolen data. A company spokesperson attributed the breach to a failure in internal back-end checks, not the AI agent itself, and noted that regulators and affected users were being notified. The incident highlights ongoing challenges as Meta accelerates its AI integration across platforms like Facebook and WhatsApp, investing billions to compete with leaders like OpenAI and Anthropic. CEO Mark Zuckerberg has framed AI as central to the company’s future, though the transition has faced internal friction, including employee resistance to activity-tracking programs and mass layoffs. Security experts warn that AI’s growing role in digital infrastructure could amplify risks, including foreign adversaries leveraging public data for influence campaigns. The breach underscores the need for updated security frameworks as AI adoption expands.