On June 4, 2026, Representatives Jay Obernolte and Lori Trahan released a discussion draft of the Great American Artificial Intelligence Act of 2026 (GAAIA). The GAAIA is the latest federal attempt at timely and coherent technology-based regulations.

This bill, however, takes a notably different approach. Narrow enough to be effective, GAAIA sidesteps the pitfalls that have plagued federal privacy law efforts by limiting scope and preemption. Not yet formally introduced, this nearly 270-page discussion draft seems to mostly target the big players (the large AI companies building foundational AI models) and is centered around transparency and innovation.

This article covers the key provisions of GAAIA, what organizations are actually in the crosshairs, and how it fits alongside the June 2nd Executive Order related to artificial intelligence.

Key Provisions of GAAIA

At a high level, GAAIA has four major areas of focus: frontier AI governance, workforce, cybersecurity, and research and development. Key provisions include:

  • Center for AI Standards and Innovation. GAAIA would formally establish the Center for AI Standards and Innovation  (CAISI) within the Department of Commerce, which would develop voluntary AI guidelines, best practices, and security standards, monitor and evaluate foreign and domestic AI system developments, support the efforts of the National Institute of Standards and Technology (“NIST”), and more.
  • Obligations on Large Frontier Developers. Organizations with annual revenues exceeding $500 million that develop frontier models (broadly, AI models trained on broad data sets and designed for generality of output, utilizing computing power greater than 1026 operations)would have certain obligations, including publishing an AI framework and transparency reports and reporting “critical safety incidents” (i.e., unauthorized access to or exfiltration of model weights, failure of risk-mitigation measures, or loss of control) to the CAISI. The published framework and reports would include disclosures related to the capabilities of, and standards, assessment procedures, risks, security, and governance applicable to, the organization’s frontier models.
  • Independent Verification Organization.  CAISI would license Independent Verification Organizations (IVOs) to audit large frontier developers’ compliance with GAAIA and assess risk mitigation in established AI frameworks.
  • AI Whistleblower Protections. Employers would be prohibited from retaliating against employees and independent contractors for lawfully reporting violations of federal AI law.
  • Federal Preemption. GAAIA would preempt state laws regulating the development of AI models for three years post-enactment.
  • Fraud. GAAIA would increase penalties when AI is used to commit fraud and add penalties when AI is used to impersonate government officials.
  • Education, Research, and Workforce. GAAIA would direct federal investment in AI literacy, expand access to AI research and education, increase insight into AI-related labor market data, and require disclosures for certain AI-related layoffs.
  • International Cooperation. GAAIA would require cooperation with other countries in the development of international AI technical standards and formation of coalitions to promote the adoption of such standards.

Who GAAIA Actually Targets

The heavy compliance obligations of GAAIA would fall on large frontier model developers, which are organizations with annual revenues exceeding $500 million that are building powerful and foundational AI models (think OpenAI and Anthropic). GAAIA would not directly impact an organization that is only deploying AI or developing models at a smaller scale, but there are indirect implications worth watching. The obligations GAAIA would place on developers would inevitably flow downstream, into vendor contracts, into the terms governing the AI tools other organizations rely on, and into the compliance expectations that sophisticated counterparties demand.

GAAIA and the Executive Order

Two days before the GAAIA discussion draft was released, President Trump signed an  executive order titled “Promoting Advanced Artificial Intelligence Innovation and Security” (the EO). This EO, in pertinent part, directs a coalition of federal officials to establish a framework for the secure deployment of frontier AI models, including a voluntary process by which developers may provide the government with early access to models prior to broader release. Read together, the EO and GAAIA reflect a federal push toward centralization of AI model development regulations, with the heaviest obligations placed on frontier model developers.

What’s Next?

GAAIA is far from law. It is still in the discussion draft phase, it has not yet been introduced and has already faced opposition. That said, the federal government clearly intends to regulate the development of AI and, at this point, it appears to be prioritizing transparency, security, and innovation. Whether GAAIA passes in its current form or not, the regulatory environment surrounding AI is not getting simpler.

We will continue to monitor updates. If you have questions about GAAIA or general AI compliance considerations, Taft’s Privacy, Security & AI attorneys are available to assist. As always, please sign up to receive emails of our latest posts here on Privacy and Data Security Insights, and follow us on LinkedIn for the latest in privacy, security, and artificial intelligence legal news.