Congress trailing on AI rules
Reporting argues US Congress is 'losing the race' to regulate AI as state bills continue to move, raising compliance and data‑residency questions for vendors and customers. The evolving legal patchwork will shape sell decisions around cloud vs on‑prem deployments. (washingtonpost.com)
State capitols introduced a surge of AI bills this cycle — MultiState counted 1,561 AI-related bills as of March 2026, up from 1,208 introduced in 2025. (multistate.ai) California has moved fastest: AB 2013 (the Generative AI Training Data Transparency Act) requires developers to publish “high‑level” summaries of training datasets effective Jan. 1, 2026. (perkinscoie.com) The state also enacted the Transparency in Frontier AI framework (SB 53/TFAIA) and the California AI Transparency Act (SB 942), which impose frontier-model and output‑labeling obligations that lawmakers set to begin implementation in 2026. (whitecase.com) Congressional attempts to block state action collapsed: a proposed 10‑year federal moratorium on state AI rules was pulled from GOP budget legislation and Senate leaders voted overwhelmingly (99–1) to strip the moratorium language. (pbs.org) Enforcement and compliance headaches are material — SB 942 authorizes civil penalties (reported as up to $5,000 per day for violations) while AB 2013’s mandated disclosures leave trade‑secret and confidentiality protections unclear. (mayerbrown.com) Market responses are already technical: hyperscalers publish data‑residency guarantees and regional hosting options (Google Cloud’s generative‑AI residency guarantees; AWS guidance on meeting residency requirements) as customers weigh hybrid or on‑prem deployments for control and auditability. (cloud.google.com) The policy gap is driving renewed federal pressure even as skepticism about Congress’s pace persists — Axios reports fresh calls for national rules after the moratorium’s failure, and The Washington Post quotes Rep. Sam Liccardo warning Congress struggles to keep up. (axios.com)
