This report finds that mandatory gene synthesis screening is a necessary, cost-effective biosecurity safeguard that closes a critical federal gap, strengthens market integrity, and can be implemented with minimal burden or risk of regulatory capture.

Research examines areas of workforce and AI policy that lawmakers should explore, including data and measurement, workforce development and education, improved social safety nets, and place-based and industry-level policies.

This report examines three policy approaches for addressing AI agent harms: the “stick” of conventional tort liability; the “carrot” of liability immunity in exchange for proactive governance measures; and the “net” of no-fault compensation schemes that provide swift remedy without fault attribution.

This report examines pathways for implementation of the Trump Administration's May 5th Executive Order, Improving the Safety and Security of Biological Research, offering policy ideas that make it more difficult for bad actors both domestically and globally to access genetic material that would allow them to produce deadly pathogens.

State-level AI legislation does not hinder public interest in and adoption of generative AI tools—in fact, it may help drive it. As policymakers weigh the risks and benefits of AI regulation, this paper analyzes AI search data and AI regulation.

Author Iskandar Haykel examines how state laws would be impacted by federal legislation that would preempt state AI regulation for the next 10 years. This research explores case studies of state tech laws in areas that include consumer protections, developer transparency, kids online safety, and more.

This report explores the gap between economist forecasts of AI's labor disruption and technologists' predictions, finding that the difference can be partially explained by failures of the General Purpose Technology framework in assess AI's impact.

ARI ranks seven big-name AI models on measures of transparency. The study explores transparency in four primary areas, including user-facing documentation, risk and safety, technical transparency, and evaluation and impact. The research uses 21 metrics to rank models in each of the four areas.