Meta's Muse AI shares user conversations due to design flaw

meta's ai assistant muse, launched recently, had a design flaw in its share links feature that made it easy to access sensitive information shared with the ai assistant. security researchers from claroc discovered 6,800 links created by muse users and found sensitive information, including confidential executive-search call notes, internal revenue reviews, loan applications, and staff vacation calendars. meta has suspended the sharing feature while reviewing the report and considering changes to how these links can be discovered. this incident highlights the importance of secure sharing features in ai products.