News & Blog

June 19, 2026—On June 4, 2026, U.S. Reps Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a 269-page bipartisan discussion draft dubbed The Great American Artificial Intelligence Act (GAAIA). The draft legislation organizes U.S. AI policy around frontier model governance, preemption of state AI laws, cybersecurity, research infrastructure, and workforce development. It is best seen as a proposal for how the government may support, test, secure, and regulate AI. The legislation draws on the Bipartisan AI Task Force report’s recommendations released at the end of the 118th Congress. Holten Stringer, associate vice president and AI policy advisor at Van Scoyoc Associates, offers a breakdown of the proposed U.S. legislation and implications for university-industry collaboration in this 3-Minute Read. Although it is a U.S.-specific issue, the policy debate will be watched by governments around the world. 

Voluntary Standards and Evaluation 

GAAIA would authorize the U.S. National Institute of Standards and Technology (NIST) Center for AI Standards and Innovation (CAISI), which is already developing voluntary guidelines, evaluation tools, and AI security standards. For covered frontier developers (those with more than $500 million in annual gross revenue), GAAIA would require publicly posted frontier AI frameworks, model disclosures, critical safety incident reporting, independent audits, and whistleblower protections. Universities and companies could see downstream impacts through procurement, vendor thoroughness, research agreements, and model citations. 

State and Local AI Law Implications 

Section 3 is predictably the most contentious topic of the discussion draft. It calls for a mandatory three-year preemption of state and local laws that specifically regulate AI model development, while preserving state laws of general applicability, common law remedies, and laws regulating AI deployment, distribution, offering, or use. 

This distinction is critical. For industry, a limited U.S. government lane for model development could reduce uncertainty created by state AI laws. For higher education, it may clarify parts of the legal environment for model training, fine-tuning, and research infrastructure. It is not, however, a blanket exemption. AI used in hiring, healthcare, education, admissions, financial aid, research administration, or consumer-facing tools may still fall under state, national, and sector-specific rules. This follows a failed congressional push for a 10-year moratorium on state AI laws that narrowly fell short of inclusion in the first budget reconciliation package. GAAIA would instead create a narrower preemption that sunsets after three years. 

Building AI Research and Workforce Pathways 

GAAIA is also a research and infrastructure bill. It would codify the National Artificial Intelligence Research Resource (NAIRR), which began as a U.S. National Science Foundation (NSF)-led pilot to expand access to computing infrastructure, data, models, software, training, and user support for researchers and educators who lack hyperscale resources. Under GAAIA, NAIRR would become a statutory resource for researchers, institutions, government entities, and small private-sector entities. 

The bill also points toward more shared AI testbeds. NIST, NSF, and the U.S. Department of Energy (DOE) would be tasked with building programs that connect national laboratories, national labs, government agencies  universities, and companies of all sizes to test, evaluate, benchmark, and assess AI systems. For those with expertise in red-teaming, interpretability, cybersecurity, energy efficiency, advanced manufacturing, and AI for science or education technology, this could create new national partnership pathways. 

GAAIA would also expand NSF authorities for AI literacy, scholarships, fellowships, and capacity building at institutions outside the top tier of U.S. R&D expenditures, including community college Centers of AI Excellence. The U.S. Department of Labor would be directed to include new AI-sensitive occupation forecasts, a workforce research hub, voluntary AI adoption reporting, survey updates, and Worker Adjustment and Retraining Notification (WARN) Act disclosures when AI is a substantial factor in a qualifying layoff. These provisions may create better labor-market signals, but they could also require more deliberate internal tracking of how AI affects job design, displacement, upskilling, and training. 

Looking Ahead 

Unsurprisingly, the bipartisan discussion draft has already drawn criticism from both sides of the aisle, with Democrats arguing it does not go far enough and Republicans arguing it goes too far. Rep. Obernolte has said he looks forward to working with House leadership to begin introducing elements of GAAIA as separate bills, allowing each committee of jurisdiction to approve its section.  

As of mid-June 2026, leaders of the U.S. House Science, Space, and Technology Committee indicated that it is ready to consider the bill’s nonregulatory research, standards, and innovation provisions. The most likely outcome is that multiple bills, mirroring parts of the discussion draft, will be introduced, and a few provisions will ultimately land in the year-end defense authorization bill. While a draft of GAAIA was reportedly backchanneled to the White House, the Trump Administration has not yet publicly endorsed or disavowed specific aspects of the proposal. 

We want to hear from you. How is your organization adapting to the evolution of AI and emerging frontier models? What specific portions of GAAIA excite or concern your team? Share your perspective on LinkedIn. 

Go Deeper

Download the full-text of the Great American AI Act discussion draft here.   

Download a section-by-section summary of the Great American AI Act discussion draft here.  

The 3-Minute Read is a UIDP member information piece and does not represent the opinions of our members or representatives. We welcome your comments on our LinkedIn profile.