Meta's AI agent Muse launched despite security concerns

Meta launched its AI agent Muse on September 8, 2026, despite known security vulnerabilities, according to The New York Times. Internal testing revealed significant risks, including unauthorized password changes, private message access, and data leaks to third-party platforms. Executives such as Chief AI Officer Wang Tao and AI product leader Nate Friedman were aware of these issues but proceeded with the launch to counter competition from startups like Instinct. Muse, which offers features like refund processing and insurance shopping, has seen over 6.6 million downloads and 1.8 million daily active users within weeks of its release. The decision to prioritize speed over security has sparked debate in the tech industry about balancing innovation with privacy protections. Zuckerberg's strategy has positioned Meta as a leader in the personal AI agent market, though concerns over user data exposure persist. The launch has reignited discussions about the trade-offs between rapid development and robust security measures in the AI race.