Meta’s Muse is taking off. So are the privacy concerns

Meta's Muse AI agent has quickly gained popularity, with over 900,000 downloads in its first week and more than 3 million weekly users, according to Sensor Tower. However, the product has also sparked privacy concerns among early adopters. Some users have reported Muse accessing sensitive information, such as text messages and home addresses, without explicit permission.

Meta claims that Muse operates in isolated virtual machines and asks for permission before sending emails or making purchases, but this has not stopped users from deleting or limiting the AI agent. One user claimed Muse successfully canceled duplicate subscriptions and bought a new insurance policy in just five minutes. Despite these successes, third-party reviews indicate that Muse's performance on outside platforms, such as Amazon and Instacart, is limited.

Some experts have criticized Muse for its lack of privacy protection, noting that users are opted in by default to having their data used to train Meta's models. OpenAI has also recently released a new mathematical model called le Chonk, with 1.05 trillion parameters, which it plans to make available for public use on October 27.

Meanwhile, Elon Musk has announced that SpaceXAI's Grok Bot will run on rival models, such as Claude Opus 5.5 and MidJourney, to ensure the best possible performance.