AI Safety
Testing, oversight, and accountability for the most capable AI systems.
AI safety is the field that studies how to build AI systems that behave reliably, predictably, and in ways that align with human intentions — especially as systems become more capable.
Why it matters to voters
Current frontier models can write code, take actions on the web, and assist with biological or cyber tasks. As capabilities grow, the cost of mistakes — accidental or deliberate — grows with them.
Recent developments
- 2025New York's RAISE Act becomes one of the first state laws requiring safety testing for frontier models.
- 2026Trump executive order on a single federal AI framework challenges state-level safety laws.
Why this mattersFederal preemption could override every state AI safety law currently on the books.
Political & policy relevance
The central fight is between a single federal framework (favored by the largest AI companies and Leading the Future) and state-by-state safety rules (favored by Anthropic-aligned PACs and many state legislators).
Candidates working on this
Alex Bores · Valerie Foushee · Josh Gottheimer
Further reading
Last updated 2025