This bill requires the Department of Homeland Security to run prize competitions aimed at improving artificial intelligence safety and reliability. Specifically, the DHS Secretary must launch at least one competition focused on AI interpretability—making AI decision-making processes understandable to humans—and at least one competition focused on adversarial robustness, which means developing AI systems that can withstand attacks or manipulation attempts. Both competitions must begin within 270 days of the bill's enactment and will involve consultation with other federal agencies, the National Institute of Standards and Technology, and industry experts. The DHS Secretary has flexibility in structuring the competitions, which may include multiple phases and different types of submissions such as frameworks, AI models, and red-team testing. The bill authorizes $10 million in funding for fiscal years 2027 through 2031, and requires the DHS Secretary to report to Congress within 180 days after the first competition concludes, evaluating the results and recommending any further action needed to advance these research fields.