Bipartisan Congress Signals End to Hands-Off AI Era With Governance Framework Push

The U.S. Congress is advancing its first comprehensive AI governance framework, with bipartisan momentum building behind the Great American AI Act and White House negotiations underway to preempt state-level AI laws. The shift is driven by a hard deadline: Colorado's comprehensive AI Act takes effect June 30, and 31 other states are watching.

For any organization deploying AI in hiring, lending, healthcare, or other high-stakes domains, the next six weeks are the most consequential period in U.S. AI regulation to date.

Why the Hands-Off Era Is Ending

Since ChatGPT launched mass AI adoption in late 2022, the federal government's approach was largely deferential — executive orders set voluntary guidelines, but no binding law reached the president's desk. Industry self-regulation filled the gap.

That changed this spring. The combination of rapid AI deployment across critical sectors, a wave of 134 AI-related state bills in 31 states, and Colorado's June 30 enforcement date made inaction increasingly untenable for congressional leaders in both parties.

The Great American AI Act — the leading federal vehicle — addresses four core areas: frontier AI development requirements, workforce transition obligations, cybersecurity standards for AI in critical infrastructure, and international cooperation frameworks with allied nations. Neither a pure industry wish list nor a sweeping regulatory overhaul, the bill reflects genuine compromise driven by the urgency of the moment.

The Preemption Calculation

Alongside the Great American AI Act, the White House is negotiating a federal preemption deal: the administration would back passage of kids-online legislation (derived from KOSA) and deepfake regulations in exchange for federal law superseding state AI governance frameworks.

For businesses, this is the most consequential element. A company running AI-powered hiring software in Colorado, Texas, and New York currently faces three diverging compliance paths. Federal preemption would replace that with a single national standard — but at the cost of giving up whatever stronger protections individual states have enacted.

The tradeoff is real. Preemption benefits companies operating across state lines. It may roll back protections in states like Colorado, which passed the most comprehensive AI law in the nation. Civil liberties groups are pushing back; industry coalitions are pushing just as hard in the other direction.

What Colorado's June 30 Law Requires

Colorado's AI Act is not waiting for Congress. Compliance is mandatory by June 30 for organizations using AI in:

• Hiring and employment decisions
• Housing and lending assessments
• Healthcare treatment or diagnosis
• Education admissions and placement

Obligations include conducting and documenting risk assessments before deployment, providing transparency notices to individuals affected by AI decisions, and maintaining bias auditing records available on request. Any company operating AI in these categories in Colorado without a compliance program in place by June 30 faces legal exposure from day one of enforcement.

What the Great American AI Act Would Add

The federal bill expands scope beyond Colorado's framework in several directions.

Frontier AI provisions require the largest models to undergo third-party safety evaluations before deployment. This is the most contested element — Anthropic and OpenAI have publicly supported binding evaluations but differ on methodology; Meta and Google have argued the requirements entrench incumbents by pricing smaller players out of compliance.

Workforce provisions would require employers to notify workers before deploying AI systems that substantially change job duties, and contribute to a federal retraining fund for displaced workers. The size of required contributions and the definition of "substantial change" are still being negotiated.

International cooperation provisions align the U.S. with G7 AI governance frameworks and establish a coordination mechanism with the EU AI Act, reducing compliance friction for multinationals operating in both markets.

How Companies Are Responding Now

Legal teams at large enterprises are broadly recommending one of three approaches:

1. Complete Colorado compliance now and treat it as the national template. Moderate cost. The risk: federal standards may diverge from Colorado's.
2. Pause AI deployments in regulated categories until federal preemption clarifies the landscape. Cost: foregone AI productivity and falling behind competitors.
3. Build dual-track compliance that satisfies Colorado today and positions for federal standards anticipated later this year. Higher upfront investment, lowest ongoing risk.

Most outside legal counsel are recommending track three for any company with significant AI deployment in covered categories. The work done for Colorado compliance is unlikely to be wasted even if federal standards arrive — the baseline obligations will be at least as stringent.

What to Watch

June 30 is the immediate marker. Colorado's enforcement starts, and the preemption negotiation enters its most active phase — because once Colorado is live, preemption becomes politically easier to sell to holdout legislators.

After that, watch for White House preemption language attached to KOSA or a standalone vehicle moving through Senate Commerce before the summer recess. If that deal stalls, expect three to five more states to pass Colorado-style laws before year end — deepening the very fragmentation a federal framework is meant to prevent.

The hands-off era didn't end with a single vote. It ended with a cascade of state action that made federal inaction untenable. Congress is responding. The timeline is now measured in weeks, not years.

By Hector Herrera