The National Security Agency (NSA) has been secretly using a banned artificial intelligence system, dubbed Mythos AI, to conduct signal intelligence on US citizens without warrants. Internal leaks reveal that the agency fed intercepted communications, metadata, and pattern-recognition datasets into the system, which the government had previously blacklisted for 'unacceptable safety and accountability risks.'
This revelation exposes federal AI bans as mere theater. By ignoring its own blacklist, the NSA has shredded any remaining accountability for surveillance programs. Unlike earlier Snowden-era revelations, this time the machine is learning as it spies—analyzing emails, calls, and location data without court oversight.
Congress has announced hearings and demands for an inspector-general probe. Other intelligence agencies are now being audited to determine if they also deployed the same prohibited tool. The story is ongoing, with significant implications for privacy and the rule of law.