New Blueprint Maps U.S. Federal Framework for Frontier AI Safety

A newly released blueprint calls on the U.S. federal government to build a durable framework for governing increasingly capable AI systems, proposing a three-part strategy that anchors federal authority in state-level frontier safety laws already moving through legislatures.

The blueprint's authors are direct about the stakes: the federal government must create institutions "capable of evolving alongside the technology itself." The document arrives at what its publishers call "an important moment to seize," with states having begun to harmonize their approaches and the White House having issued a new executive order on Promoting Advanced Artificial Intelligence Innovation and Security.

Why does this matter now? Frontier AI development has outpaced the patchwork of voluntary commitments and sector-specific rules that currently govern it in the United States. The blueprint argues that the country now has both political momentum — in the form of state legislation — and an institutional starting point to convert that momentum into a permanent federal structure.

What does the blueprint actually propose?

The document lays out three pillars.

• A national framework built on state consensus. The blueprint proposes leveraging "the emerging consensus reflected in state frontier safety laws" as the foundation for federal legislation, rather than starting from scratch.
• Strengthening CAISI. The blueprint identifies CAISI — the federal government's Center for AI Safety and Security — as the primary institution for frontier AI safety and calls for reinforcing it as the central node of federal oversight.
• A government-wide resilience plan. The third pillar mobilizes agencies across the federal government to address the national security and public safety challenges that frontier AI systems pose.

The three-part structure is deliberate. State laws supply political legitimacy and tested policy language. CAISI supplies institutional capacity. The resilience plan addresses the reality that frontier AI risks — from biosecurity concerns to critical infrastructure vulnerabilities — do not sit neatly within any single agency's jurisdiction.

Which state laws does the blueprint point to?

The blueprint names three specific pieces of legislation as evidence that a harmonized state-level approach is already forming.

• California's SB 53, the state's frontier AI safety legislation.
• New York's RAISE Act, a companion effort in the Empire State.
• Illinois's SB 315, the Midwest entry in what the blueprint describes as a converging set of state frameworks.

This state-level activity matters for a specific reason. When multiple large states pass comparable frontier safety laws, developers face de facto national standards regardless of federal action. The blueprint's argument is that this emerging consensus gives Congress a ready-made policy baseline — one that has already survived legislative debate in multiple jurisdictions — reducing the friction that has stalled previous federal AI safety efforts.

The authors frame the state laws not as a substitute for federal action but as a foundation for it. "The federal government must now build on that foundation," the blueprint states, "and create a durable federal framework capable of evolving alongside the technology itself."

Where does the White House executive order fit?

The blueprint treats the White House's new executive order on Promoting Advanced Artificial Intelligence Innovation and Security as "another important step forward" in the governance picture. The order, issued by the current administration, complements the legislative track running through the states.

The sequencing matters. Executive orders can be revoked by subsequent administrations. Legislation, once enacted, is far harder to unwind. By pairing the executive order's near-term measures with a push for durable statutory authority grounded in state-level consensus, the blueprint aims to make frontier AI governance resistant to political turnover.

That durability theme runs throughout the document. The word choice is consistent: "durable federal framework," "durable institutions." The blueprint's central claim is that the United States does not merely need AI rules — it needs rules and institutions designed to adapt as frontier capabilities advance, rather than rules that are obsolete on arrival.

What is the role of CAISI?

The second pillar of the blueprint gives CAISI — the federal government's primary institution for frontier AI safety — an expanded, stabilized role. The document calls for "strengthening CAISI as the U.S. federal government's primary institution for frontier AI safety."

The emphasis on one lead institution reflects a common critique of U.S. AI governance to date: responsibility for frontier AI safety has been distributed across multiple bodies with overlapping mandates and no clear authority. The blueprint's answer is consolidation around CAISI, backed by the national framework described in the first pillar.

Why a whole-of-government resilience plan?

The blueprint's third pillar — the "broader resilience plan across government" — addresses the national security and public safety challenges posed by frontier AI. The framing acknowledges that a single safety agency cannot manage risks that span defense, homeland security, public health, and critical infrastructure.

The resilience plan is the blueprint's mechanism for distributing readiness across agencies while keeping CAISI as the coordinating center for frontier AI safety specifically.

What happens next?

The blueprint is a proposal, not legislation. Its immediate practical effect is to give policymakers a concrete starting text: a three-part structure, a named set of state laws to draw from, a designated lead institution, and an executive order to build upon.

The document's own closing logic points to the path forward. States have legislated. The White House has acted by executive order. The remaining step, in the blueprint's words, is for the federal government to "build on that foundation and create a durable federal framework capable of evolving alongside the technology itself."

Readers can consult the full blueprint, linked by its publishers alongside the announcement, for the complete policy detail.