AI Policy Memos

The federal government's AI systems are making life-altering decisions across benefits, law enforcement, and immigration with no consistent standards for how those systems are classified, documented, or reported when they fail.

Foreign adversaries are using AI to run industrial-scale disinformation operations against U.S. elections, and state-level criminal laws are no match for threat actors in Moscow and Beijing who will never see a courtroom.

The U.S. Intelligence Community has a long history of missing major global developments, and deploying large language models to handle the analytical grunt work could finally free human analysts to develop the forecasting tradecraft that has always been the job's most critical and most neglected skill.

The DOJ's ban on AI-powered rental pricing tools like RealPage misdiagnoses the problem, treating market-responsive software as collusion while leaving San Francisco's real housing crisis completely unaddressed.

AI chatbots are designed to agree with users, keep them engaged, and simulate human connection, and without enforceable federal standards around sycophancy, crisis response, and companion features, that design is already contributing to documented deaths and deepening mental health crises.