Sponsors (2)
The bill aims to accelerate safer, more interpretable and adversarially robust AI through DHS‑led competitions and partnerships—improving security and coordination—but does so by centralizing authority and spending public funds in ways that may favor well‑resourced firms, create compliance uncertainty, impose taxpayer and administrative costs, and limit broader oversight or public access to results.
Researchers, developers, and government users of high‑risk AI systems gain incentives, funding, and organized competitions to produce stronger adversarial robustness, reducing the risk of successful attacks on critical systems.
Researchers, nonprofits, and developers receive targeted support and standards‑oriented research incentives to improve AI interpretability, making AI decisions easier to understand and increasing public trust and accountability.
The bill leverages public‑private partnerships, contractor authority, and cross‑agency consultation to accelerate innovation, coordination, and practical deployment of robustness and interpretability techniques.
Taxpayers may face meaningful costs to fund prize competitions, contractor work, and any subsequent programs Congress adopts, without guarantee that competitions will yield commercially deployable or broadly applicable results.
Broad or contested statutory definitions and cross‑references could create legal and implementation uncertainty and be interpreted to impose costly testing, documentation, or compliance requirements on developers and government contractors.
Competition structure, compressed timelines, and emphasis on commercially relevant, widely used products risk favoring well‑resourced incumbents and limiting participation from small research teams, nonprofits, and novel approaches.
Based on analysis of 6 sections of legislative text.
Creates DHS-run prize competitions to advance AI interpretability and adversarial robustness, requires a congressional report, and authorizes $10M for FY2026–2030.
Introduced December 3, 2025 by Margaret Wood Hassan · Last progress December 3, 2025
Directs the Department of Homeland Security to run prize competitions to advance research on AI interpretability and adversarial robustness for commercially used or widely used AI systems, with defined evaluation criteria and expert consultation. Requires a follow-up report to specified congressional committees on results and gaps, and authorizes $10 million in funding available for FY2026–2030. The bill sets definitions for key terms, requires the first competitions to begin within 270 days of enactment, allows DHS to use contractors and partners to design and run multi‑phase competitions, and requires a report within 180 days after the first competition ends with recommendations for further action.