On 20 March 2026 the White House published a national artificial intelligence legislative framework. In its own framing, it issues a “comprehensive national legislative framework” addressing the most pressing AI policy topics, and provides recommendations for Congress to create a federal standard.
The framework’s objectives include enabling innovation and maintaining American leadership in AI, child safety, pro-growth community-centred development, respect for intellectual property rights, prevention of censorship, and building an AI-ready workforce.
Why a “framework” rather than a bill
US AI governance has long been characterised by federal absence and state-first action. Several states have legislated on their own, leaving companies with a fragmented compliance environment. The value of the White House framework is not that it has legal force, but that it seeks to consolidate the federal legislative agenda onto a single list.
For companies, the key variable ahead is whether Congress will turn it into a uniform standard, and how that standard will interface with existing state rules.
Three implications for companies
First, intellectual property is expressly on the agenda. This echoes the US Supreme Court’s denial of certiorari in Thaler v. Perlmutter in the same period — the protectability of AI-generated content and fair use of training data are becoming core US AI policy issues. See our note on the Thaler decision.
Second, incident-reporting requirements are tightening. Alongside the framework, federal-level arrangements for reporting AI safety incidents are advancing. Companies should build incident identification and reporting capability into their products — see our note on the US AI model review executive order.
Third, state rules have not disappeared. Federal progress will not immediately displace state law. Until a uniform federal standard lands, companies should still design to the strictest state requirements.
How to respond if you operate in both China and the US
First, build a compliance documentation system that can be reused. Technical documentation, risk management and data governance should satisfy Chinese filing requirements while moving towards internationally conventional formulations — see our multi-jurisdiction service.
Second, treat intellectual property and data provenance as asset questions. Training-data provenance, ownership of generated content and model ownership have become due-diligence focal points in transactions and financing in both China and the US.
Third, track rather than predict. US AI legislation remains in flux; rather than betting on one direction, build a standing tracking mechanism — see our note on the Manus foreign-investment security review and our EU AI Act service.
The White House framework is an agenda-setting move. Its significance is that it turns “will the US have uniform AI rules” from an open question into a question of timing.
This article is based on the White House public document of 20 March 2026 and public reporting. It is for general reference only and does not constitute legal advice.