The Great American AI Act: What the Obernolte-Trahan Draft Bill Means for Copyright, Innovation, and You

The Great American AI Act: What the Obernolte-Trahan Draft Bill Means for Copyright, Innovation, and You

Published: June 5, 2026 | Category: Regulation | Read time: 8 min

On June 4, 2026, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a 269-page discussion draft of the Great American Artificial Intelligence Act — the most ambitious attempt yet to create a unified federal AI framework in the United States. The bill arrives at a moment when the patchwork of state-level AI laws has created significant compliance headaches for developers, and when copyright battles over AI training data remain unresolved in courts across the country.

But here's what makes this bill different from the dozens of AI proposals floating around Capitol Hill: it's bipartisan, comprehensive, and explicitly designed to preempt state AI laws for three years while a national standard takes shape.

For creators, developers, and businesses watching the AI copyright landscape, this bill could reshape everything — from how training data must be disclosed to who gets sued when an AI model infringes.

Let's break down what's actually in the draft, what it means for copyright, and why this conversation is happening now.

Why Now? The Patchwork Problem

If you've been following AI regulation in the US, you already know it's a mess. States have been racing to fill the federal vacuum:

• California passed AB 2013, which requires AI developers to disclose their training data sources — a law now in force with major copyright implications.
• New York enacted its own frontier AI framework, adding another layer of requirements.
• Illinois moved forward with transparency and auditing mandates.
• Multiple other states have bills in various stages of development.

The result? A company building an AI model has to navigate dozens of different legal regimes — each with its own definitions, requirements, and penalties. As Obernolte and Trahan put it in their Bloomberg Law op-ed: "Protections that depend on your zip code are not enough."

This fragmentation is especially acute for copyright. California's AB 2013 requires training data transparency; other states don't. The EU AI Act already mandates copyright transparency for models used in the European market. Meanwhile, the US Copyright Office's Part 3 report on generative AI training — released in pre-publication form in May 2025 — provided analysis but no binding rules.

The Great American AI Act aims to replace this chaos with one national standard.

What's Actually in the Draft Bill?

At 269 pages, the discussion draft is substantial. Here are the key provisions that matter most for copyright and creative industries:

1. Safety and Transparency Requirements for Frontier AI

The bill requires large frontier AI developers to:

• Publish and follow safety plans for managing catastrophic risks
• Report serious safety incidents to regulators
• Submit to audits by independent verification organizations
• Face penalties for noncompliance

For copyright, the transparency requirements are significant. While the draft doesn't go as far as the EU AI Act's detailed training data summary requirement, it establishes a structure that could evolve to include copyright-specific disclosures.

2. Preemption of State AI Laws (Three-Year Window)

This is the headline feature. The bill would preempt state AI laws for three years while the national framework takes effect. During this period:

• Companies wouldn't need to comply with California's AB 2013 (if they comply with the federal standard)
• One set of rules would apply across all 50 states
• State attorneys general could still opt in to enforcement — creating a hybrid federal-state enforcement model

This is a big deal for AI companies currently navigating the California-New York-Illinois triangle. But it also means the federal standard better be good — because it would temporarily replace all those state-level protections.

3. AI Labor Market Impact and Worker Protections

The bill requires:

• Better federal data collection on AI's labor market impact
• Improved forecasting for occupations most likely to be affected
• Additional transparency when AI is a "substantial factor" in qualifying mass layoffs

For creative workers — writers, artists, musicians, voice actors — this data could be crucial in demonstrating how AI is reshaping their industries. It also connects to the ongoing conversation about SAG-AFTRA's AI protections and the broader push for creator rights in an AI-driven economy.

4. National AI Education and Workforce Strategy

The framework invests in AI literacy, supports educators, strengthens technical education, and aims to expand AI workforce development beyond traditional tech hubs. For the copyright world, this signals that policymakers are thinking about the human creators who need to adapt — not just the companies building the models.

5. International AI Leadership

The bill strengthens America's role in shaping international AI standards, including coordination with like-minded partners. This matters for copyright because AI training data and model deployment are inherently global — a US framework that aligns with (or diverges from) the EU AI Act will shape the global playing field.

What the Bill Doesn't Do (Yet) — The Copyright Gaps

For a site called AI Copyright Legal, this is the important part. The draft bill, as released on June 4, is a discussion draft — meaning it's intended to start a conversation, not be the final word. Several copyright-critical issues are either absent or underdeveloped:

No Training Data Copyright Rule

The bill doesn't directly address the biggest open question in AI copyright law: is training on copyrighted works fair use? That question remains with the courts, where cases like NYT v. OpenAI and the various publisher lawsuits against Meta are still being litigated.

No Opt-Out Mechanism (Yet)

Unlike the EU AI Act, which includes a text-and-data-mining opt-out, the draft doesn't establish a clear mechanism for creators to opt out of having their work used for AI training. This is a gap that creator advocacy groups may push to fill during the discussion period.

No Copyright Office Role Defined

The US Copyright Office has been the primary federal entity analyzing AI copyright issues (producing three major reports since 2024). The draft bill doesn't explicitly define the Copyright Office's role under the new framework — leaving open the question of whether copyright-specific AI regulation would flow through the Copyright Office or a new AI oversight body.

Silence on AI-Generated Content Copyrightability

The bill doesn't touch the question of whether AI-generated works can be copyrighted. That issue remains governed by Copyright Office policy (the human authorship requirement) and the Copyright Office's Part 2 report.

The Political Landscape: Can This Actually Pass?

Let's be realistic about the odds. A 269-page AI bill in an election year? That's a heavy lift.

However, several factors make this different from previous failed attempts:

1. Bipartisan sponsorship. Obernolte (Republican) and Trahan (Democrat) co-authoring gives the bill credibility on both sides of the aisle.

2. Industry pressure. Companies are begging for federal preemption. Navigating California + New York + Illinois + EU rules simultaneously is genuinely unsustainable — especially for startups and mid-size companies that can't afford compliance armies.

3. The Trump executive order. On the same day the draft bill was released, President Trump signed an executive order on AI security that calls for voluntary 30-day review of frontier models — but stopped short of mandatory requirements. The bill fills the gap that the executive order deliberately left open.

4. State momentum. The more states that pass their own AI laws, the louder the calls for federal preemption become. Several states are actively moving bills — creating a "now or never" dynamic for federal action.

5. International pressure. The EU AI Act is already in force. Other nations are moving. The US risks ceding AI governance leadership if it doesn't act.

That said, the bill faces significant hurdles: an election-year Congress with limited bandwidth, potential opposition from both tech-optimists (who may see it as too restrictive) and safety advocates (who may see it as not restrictive enough), and the complexity of reconciling 269 pages of provisions.

What This Means for You (By Stakeholder)

For AI Developers

If this bill passes (in some form), you'd face:

• One national standard instead of 50 state regimes (+ EU)
• Safety plans, incident reporting, and independent audits
• Potential penalties for noncompliance
• Three years of state-law preemption to adapt

The compliance burden doesn't disappear — it shifts from multiple state regimes to a single federal one. Whether that's a net win depends on how demanding the federal standard turns out to be.

For Creators and Copyright Holders

The bill is a mixed bag:

• Good: Better data on AI's impact on creative jobs could support future protections
• Concerning: Preemption of state laws like California AB 2013 (which currently provides training data transparency) would only be acceptable if the federal standard is at least as strong
• Uncertain: The bill doesn't directly address training data copyright or fair use — meaning these battles continue in the courts

For Businesses Using AI

If you're a business deploying AI tools (not building frontier models), the bill may affect you indirectly:

• AI vendors you rely on would face new safety and transparency requirements
• Better data on AI workforce impact could affect hiring and compliance strategies
• A clear national framework could make it easier to draft corporate AI policies

For the General Public

The bill promises:

• Core AI protections extended to all 50 states (not just those with active legislatures)
• National AI literacy and workforce development investments
• A unified approach to AI safety that doesn't depend on your zip code

How This Connects to Existing Laws

The Great American AI Act doesn't exist in a vacuum. It would interact with:

| Law/Framework | Status | Interaction with Draft Bill |

|--------------|--------|---------------------------|

| California AB 2013 | In force | Would be preempted for 3 years if federal standard applies |

| EU AI Act | In force | Federal framework could align with (or diverge from) EU requirements |

| Trump AI Executive Order | Signed June 4, 2026 | Voluntary review; bill adds mandatory requirements |

| US Copyright Office AI Reports | Parts 1-3 published | Bill doesn't incorporate Copyright Office findings directly |

| State AI Laws (NY, IL, etc.) | Various stages | All would be preempted during 3-year window |

Key Takeaways

1. The Great American AI Act is a discussion draft, not a final bill. It's designed to start a national conversation — expect significant changes before any vote.

2. Copyright is largely untouched. Training data fair use, AI-generated work copyrightability, and creator opt-out mechanisms are not addressed in the current draft. These issues remain with the courts and the Copyright Office.

3. Federal preemption is the centerpiece. If passed, state AI laws (including California AB 2013) would be preempted for three years. Whether this is good or bad depends on the strength of the federal standard.

4. Bipartisan support is real but not a guarantee of passage. Obernolte and Trahan have credibility, but an election-year Congress faces enormous competing priorities.

5. The conversation is just starting. The draft bill invites feedback from workers, researchers, startups, frontier labs, educators, civil society, state leaders, and the public. This is the moment to make your voice heard.

What Comes Next?

The draft bill will now go through:

1. Public comment period — stakeholders weigh in on the 269-page framework

2. Committee hearings — the House Energy and Commerce Committee will hold hearings

3. Markup and amendments — expect significant changes

4. Potential floor vote — timing uncertain, especially in an election year

We'll be tracking this closely. The Great American AI Act could be the most consequential AI legislation in US history — or it could join the long list of ambitious tech bills that never made it to a vote.

---

This article was last updated on June 5, 2026.

Disclaimer: This article is for informational purposes only and does not constitute legal advice. AI copyright law is evolving rapidly. Consult a qualified attorney for guidance on your specific situation.