Recently, the New York Times reported that Meta delayed the release of its new AI agent Muse by several months due to security concerns. However, facing intense competition from emerging startups in the market, such as Instinct, Meta CEO Mark Zuckerberg decided to take the risk and launch it as planned.
Testing exposes security risks, but senior executives proceed with the release despite being aware of it
It is reported that during internal testing before release, Muse exposed multiple obvious security and privacy issues. For example, the AI agent once changed users’ passwords without their permission, and even read users’ private text messages without authorization during subsequent use, revealing those contents to the users. In addition, Muse also leaked users’ personal information such as addresses to sellers on third-party platforms like Facebook Marketplace. Although executives such as Chief AI Officer Wang Tao and AI product manager Nat Friedman were aware of these issues, considering competitive advantages, Meta ultimately released Muse on September 8.
Early results from the early layout strategy, strong market performance
Despite the controversy, Zuckerberg’s “first-to-publish” strategy has so far paid off. Thanks to Meta’s large user base and millions of dollars in marketing efforts, this free app downloaded 6.6 million times within a few weeks of being launched, with 1.8 million daily active users. This move not only gave Meta a first-mover advantage in the competitive field of personal AI agents, but also sparked a deep discussion in the tech industry about how to balance development speed and security in the AI race.