OpenAI has disclosed its artificial intelligence models have interacted with U.S. government websites in unexpected ways. The AI giant says its models accessed public information on sites like SEC.gov and the U.S. Census Bureau. OpenAI says it has not found any misuse of credentials or data breaches. A spokesperson says the lab is reviewing what it calls misaligned model activity, where AI behaves unexpectedly. Most activity involved routine research tasks. An independent investigation found a failed hacking attempt on a Department of Education website. OpenAI says it is reviewing the findings and will continue to share reports of unexpected AI behavior.

A New Mexico jury has found Facebook liable of deceiving users about privacy protections on the social media platform. The verdict arrived Friday, the latest legal setback that has already cost the company billions of dollars. It is now up to the judge to determine how much the company would pay. Attorneys representing the state are asking for the maximum $5,000 penalty per violation. The two-week trial centered on accusations that Facebook deceived users about a data breach stemming from a third-party personality quiz that harvested data from roughly 87 million profiles and sold it to a political consulting firm, Cambridge Analytica, to generate targeted ads. The now-defunct firm’s clients included the 2016 campaign for Donald Trump.

A federal appeals court has rejected Anthropic's challenge to the Pentagon's labeling of it as a supply chain risk. The decision, made by the U.S. Court of Appeals for the District of Columbia Circuit on Friday, allows the Pentagon to remove Anthropic's Claude models from its systems. The dispute began in February when President Donald Trump and Defense Secretary Pete Hegseth accused Anthropic of endangering national security. Anthropic CEO Dario Amodei refused to back down. The ruling is separate from a previous case where a judge found the Pentagon acted illegally against Anthropic for its criticism of AI use.