Meta Pauses Muse Chat Sharing After Researchers Expose Private Data Leak
Meta has suspended a feature in its experimental AI assistant Muse after security researchers discovered a significant data exposure. The company found that the chat sharing functionality allowed access to approximately 6,800 shared conversations, which included internal company documents and potentially sensitive user information.
The vulnerability was identified by researchers who were able to access conversations that users had chosen to share. This incident highlights ongoing concerns about data handling in AI products, particularly those featuring social or sharing components.
Meta has not yet provided a timeline for when the feature might be restored or what specific changes will be implemented to prevent similar exposures in the future. The company is reportedly reviewing its security practices across its AI product lineup as a result of this discovery.