Meta Confirms Muse AI Will Expose Its Filesystem to Users

Meta’s Muse, an AI chatbot developed by Meta Superintelligence Labs, has been found to reveal its internal file system when prompted. Initially discovered after users prodded the system, the files provided a behind‑the‑scenes look at the chatbot’s architecture, showing details that Muse had previously warned would not be shared. Meta’s Nat Friedman later clarified that this exposure is “intended behavior.” In a follow‑up X post, David Singleton of Meta Superintelligence Labs explained the choice as deliberate, indicating the company’s decision to make the filesystem accessible on request. The move raises questions about transparency, security, and user trust in AI systems.