US Comprehensive AI Framework Bill

Overview

The Great American AI Act (GAAIA) represents a landmark bipartisan legislative proposal in the United States, introduced as a discussion draft on June 4, 2026, by Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA). This comprehensive bill aims to establish a unified federal framework for artificial intelligence regulation, addressing the rapidly evolving landscape of AI technologies and their societal impact. The primary motivation behind GAAIA is to foster national uniformity in AI governance, preventing a fragmented patchwork of state-level regulations that could hinder innovation and create compliance complexities for businesses operating across state lines. The Act seeks to balance the promotion of American leadership in AI innovation with robust safeguards for public safety, national security, and individual rights. It is designed to be a foundational piece of legislation, setting broad principles and establishing institutional mechanisms for ongoing oversight and adaptation.

Key objectives of the Great American AI Act include enhancing transparency and accountability in the development and deployment of advanced AI systems, particularly 'frontier models.' It proposes the formal establishment and empowerment of the Center for AI Standards and Innovation (CAISI) within the Department of Commerce, positioning it as a central authority for AI evaluation, standards development, and incident response. Furthermore, the Act addresses critical areas such as workforce impacts, cybersecurity, and the deterrence of AI-enabled fraud. By introducing provisions for federal preemption of certain state AI laws, the GAAIA endeavors to create a more coherent regulatory environment, while also ensuring that essential state-level protections and common law remedies remain intact. The discussion draft signals a significant step towards a harmonized national approach to AI, reflecting a consensus among lawmakers on the necessity of proactive governance in this transformative technological domain.

Definitions

The Great American AI Act establishes a precise glossary of terms crucial for its interpretation and application, ensuring clarity across its various provisions. Central to the Act are definitions pertaining to different categories of AI systems and their developers. An 'artificial intelligence model' is broadly defined to encompass any computational model that, through learning or reasoning, can perform tasks typically associated with human intelligence, such as perception, reasoning, learning, or interaction. A 'frontier model' refers to the most advanced and powerful AI systems, characterized by their significant capabilities and potential for systemic impact, often requiring substantial computational resources for training and deployment. A 'large frontier developer' is designated based on specific thresholds, typically involving revenue generation exceeding a certain amount (e.g., $500 million), indicating their substantial influence and capacity within the AI ecosystem. This distinction is critical as the Act often imposes more stringent obligations on these larger entities.

Other pivotal definitions include 'catastrophic risk,' which refers to potential harms from AI systems that could lead to widespread severe negative consequences, such as critical infrastructure failure, mass casualties, or significant economic disruption. The Act also defines 'independent verification organization (IVO)' as an accredited third-party entity responsible for conducting impartial audits and evaluations of AI models, particularly frontier models, to assess their compliance with safety, security, and transparency requirements. The 'Center for AI Standards and Innovation (CAISI)' is formally defined as the federal entity established within the Department of Commerce, tasked with developing voluntary guidelines, best practices, and standards for AI, as well as overseeing the IVO licensing regime. These definitions collectively lay the groundwork for a structured regulatory approach, enabling targeted interventions and clear responsibilities for various stakeholders involved in the AI lifecycle.

Governance and Institutional Framework

The Great American AI Act fundamentally reshapes the institutional landscape for AI governance in the United States by formally establishing and significantly empowering the Center for AI Standards and Innovation (CAISI) within the Department of Commerce. CAISI is envisioned as the cornerstone of federal AI oversight, tasked with a broad mandate that includes developing voluntary guidelines, best practices, and technical standards for AI security, interpretability, and supply chain integrity. Its responsibilities extend to evaluating AI systems, monitoring technological progress, and supporting the creation of tools for synthetic content detection. The Act authorizes substantial annual funding for CAISI, recognizing the critical need for robust federal capacity in AI expertise and infrastructure. This includes granting CAISI special authorities to attract and retain top technical talent by offering competitive compensation packages, ensuring the agency can effectively fulfill its complex mission. CAISI will also play a pivotal role in administering the independent verification organization (IVO) licensing regime, accrediting and overseeing third-party auditors responsible for assessing AI model compliance.

While establishing CAISI as a central hub, the Act deliberately avoids creating an entirely new, overarching federal AI regulator. Instead, it advocates for leveraging the expertise of existing sector-specific agencies to oversee AI applications within their respective domains. This approach aims to integrate AI governance into established regulatory frameworks, allowing agencies with deep subject matter knowledge (e.g., in healthcare, finance, or transportation) to address AI-specific risks pertinent to their sectors. The Act promotes a collaborative federal ecosystem, where CAISI provides foundational standards and technical guidance, while other agencies apply these principles in a context-specific manner. This distributed governance model is intended to foster agility and responsiveness, enabling tailored regulatory responses to the diverse applications of AI across the economy. Furthermore, the Act emphasizes ongoing coordination through interagency committees, ensuring a cohesive and harmonized federal strategy for AI research, development, and regulation.

Key Focus Areas

The Great American AI Act prioritizes several key areas to establish a robust and balanced federal AI framework. A central focus is on enhancing transparency and disclosure for frontier AI models. The Act mandates that large frontier developers, defined by specific revenue thresholds, must develop, implement, and publicly post a 'frontier AI framework.' This framework must detail their risk thresholds, assessment procedures for catastrophic risks, model weight cybersecurity, and decisions regarding both internal and external deployments. Before or concurrently with deploying any new frontier model, developers are required to publish a comprehensive report disclosing the model's release date, supported languages, output modalities, intended use, any restrictions, detailed risk assessments, and mitigation steps. While redactions are permitted for trade secrets, cybersecurity, public safety, or national security, the overarching goal is to provide greater insight into the capabilities and potential risks of advanced AI systems to regulators and the public.

Another critical area addressed by GAAIA is risk management and safety. The Act directs CAISI to advance collaborative frameworks, standards, and guidelines for mitigating risks associated with AI systems. This includes supporting the development of technical standards and guidelines for testing AI systems for bias and promoting trustworthy AI. The legislation also tackles federal preemption of state laws, aiming to create a uniform national standard by preempting any state or local law specifically regulating the development of AI models. However, this preemption is carefully delineated, explicitly not affecting laws of general applicability, common law remedies, or laws regulating AI use or deployment. This ensures that states can continue to protect consumers and address specific harms without impeding the innovation cycle at the development stage. Furthermore, the Act includes provisions for workforce impact, requiring additional disclosures under the WARN Act when AI is a substantial factor in mass layoffs, and calls for improved federal data collection and forecasting regarding AI's effects on employment. It also strengthens cybersecurity through amendments to the Cybersecurity Act of 2015 and increases penalties for AI-enabled fraud, alongside robust whistleblower protections for individuals reporting violations of federal AI law, underscoring a holistic approach to responsible AI development and deployment.

Implementation Framework

The implementation framework of the Great American AI Act is designed to be multi-faceted, relying on a combination of federal oversight, industry self-regulation guided by government standards, and independent third-party verification. At its core, the Center for AI Standards and Innovation (CAISI) is tasked with developing and disseminating voluntary guidelines, best practices, and technical standards that will serve as the foundational benchmarks for AI development and deployment across various sectors. These guidelines will cover critical aspects such as AI security, adversarial robustness, interpretability, supply chain threats, and model tampering. CAISI's role is not merely advisory; it will actively engage in evaluating AI systems and monitoring the progress of AI technologies, providing crucial data and insights to inform policy decisions and regulatory updates. This proactive approach ensures that the federal government remains responsive to the rapid advancements in AI, adapting its guidance to new challenges and opportunities as they emerge.

A significant component of the implementation strategy is the establishment and oversight of the Independent Verification Organization (IVO) licensing regime. Under this regime, CAISI will accredit and regulate third-party entities—the IVOs—that are responsible for conducting independent audits and evaluations of AI models, particularly frontier models. These IVOs will assess compliance with the transparency, safety, and risk mitigation requirements outlined in the Act. Developers of frontier AI models will be required to engage these licensed IVOs to perform audits before or concurrently with deployment, ensuring an objective assessment of their systems' adherence to federal standards. The Act grants IVOs necessary access to company materials for their audits and provides them with immunity from claims of loss stemming from an AI model they have audited, thereby encouraging thorough and unbiased evaluations. This framework aims to foster a culture of accountability and continuous improvement within the AI industry, leveraging external expertise to bolster trust and safety in AI systems.

Monitoring and Evaluation

Monitoring and evaluation under the Great American AI Act are critical components designed to ensure the ongoing effectiveness, adaptability, and compliance of AI systems with the established federal framework. The Center for AI Standards and Innovation (CAISI) is assigned a central role in this process, tasked with continuously evaluating AI systems and monitoring the overall progress of AI technology. This involves conducting research, developing testing methodologies, and collecting data on AI performance, risks, and societal impacts. CAISI's monitoring activities will inform policy adjustments, update technical standards, and provide the federal government with a comprehensive understanding of the AI landscape. The agency will also be responsible for tracking incidents involving AI systems, particularly those related to catastrophic risks or critical safety failures, enabling a rapid and coordinated response to emerging threats. This continuous feedback loop is essential for a regulatory framework dealing with a technology as dynamic as artificial intelligence.

A key mechanism for evaluation is the Independent Verification Organization (IVO) regime. IVOs, accredited and overseen by CAISI, will conduct mandatory, independent, third-party audits of frontier AI models. These audits will assess developers' compliance with their publicly posted frontier AI frameworks, including risk thresholds, mitigation strategies, and cybersecurity measures. IVOs will report their findings directly to CAISI, and state Attorneys General who opt-in will also receive these audit and assessment reports, facilitating a broader oversight network. The Act mandates that these evaluations be thorough, objective, and conducted by entities free from conflicts of interest. The insights gained from these independent verifications will be crucial for identifying systemic issues, validating compliance efforts, and ensuring that AI systems meet the safety and transparency benchmarks set by the Act. This dual approach of federal monitoring by CAISI and independent third-party evaluation by IVOs creates a robust system for accountability and continuous improvement in AI governance.

Penalties, Liability, and Appeals

The Great American AI Act includes specific provisions for penalties and liability, designed to ensure compliance and deter misuse of AI technologies. For instances of fraud-related offenses where AI is utilized, the Act significantly increases maximum fines for federal mail fraud, wire fraud, bank fraud, and money laundering statutes, raising them from $1 million to $2 million. This enhancement reflects a broader enforcement trend to apply existing fraud and misconduct frameworks to AI-enabled conduct, recognizing the potential for AI to amplify the scale and sophistication of such crimes. Additionally, the Act introduces new penalties for AI impersonation of federal officials, underscoring the government's commitment to safeguarding against deceptive AI applications that could undermine public trust and governmental integrity. These measures aim to provide strong disincentives against malicious or negligent use of AI in financial and official contexts.

Furthermore, the Act incorporates robust anti-retaliation protections for AI whistleblowers. It explicitly prohibits AI companies from discriminating or retaliating against employees or independent contractors who lawfully report violations of federal AI law. If such retaliation occurs, affected individuals are entitled to significant redress, including reinstatement to their position, two times back pay with interest, compensatory damages for any harm suffered, and reimbursement for litigation costs and attorney's fees. This provision is crucial for encouraging internal reporting of non-compliance or unsafe AI practices, fostering a culture of accountability within AI development organizations. Regarding liability for Independent Verification Organizations (IVOs), the Act grants them immunity from claims of loss arising from an AI model they have audited. This immunity is intended to protect IVOs from undue legal burdens, allowing them to conduct thorough and honest assessments without fear of being held liable for the ultimate performance or failure of the AI systems they evaluate, thereby promoting their independent and objective functioning within the regulatory framework.

Relationship to Other Instruments

The Great American AI Act is designed to integrate into the existing complex tapestry of U.S. federal and state laws, while also asserting a clear federal leadership role in AI regulation. A significant aspect of its design is the proposed federal preemption of state and local laws specifically regulating the development of AI models. This provision aims to prevent a fragmented regulatory landscape across the 50 states that could stifle innovation and create compliance burdens for developers operating nationally. However, this preemption is carefully delimited; the Act explicitly states that it does not preempt laws of general applicability, common law remedies, or laws regulating the use or deployment of AI. This means that existing state laws related to consumer protection, privacy (such as the California Consumer Privacy Act), employment, healthcare, and financial services, as well as common law claims for negligence or product liability, would largely remain unaffected, ensuring continued protections for individuals and businesses in these domains. The preemption is also subject to a three-year sunset clause, allowing for re-evaluation and potential adjustments based on evolving circumstances and legislative consensus.

The Act also builds upon and interacts with other federal initiatives and existing legislation. It codifies and expands the National Artificial Intelligence Initiative Act of 2020 by formally establishing CAISI and providing it with enhanced funding and authorities, thereby strengthening the federal government's capacity for AI research, development, and policy coordination. Furthermore, GAAIA incorporates amendments to the Cybersecurity Act of 2015, extending information sharing authorities and bolstering cybersecurity efforts in response to advanced AI capabilities. It also aligns with the principles of the National Institute of Standards and Technology (NIST) AI Risk Management Framework, which provides voluntary guidance for managing AI risks, by directing CAISI to inform any revisions or additions to this framework. While the Act establishes a federal standard, it explicitly exempts state laws related to child safety, data center infrastructure, and state government procurement of AI systems from preemption, acknowledging legitimate state interests in these specific areas. This intricate relationship aims to create a cohesive national strategy that leverages existing legal structures while introducing targeted federal oversight for AI.

International Alignment

The Great American AI Act recognizes the global nature of artificial intelligence development and deployment, emphasizing the importance of international alignment and cooperation to ensure U.S. leadership and promote trustworthy AI worldwide. The Act explicitly supports American-led technical standards, aiming to influence global norms and best practices for AI safety, security, and ethical development. By investing in CAISI and its role in developing robust standards, the U.S. seeks to set a benchmark that can be adopted or referenced by international partners, fostering interoperability and reducing regulatory fragmentation across borders. This approach is rooted in the understanding that AI risks and benefits transcend national boundaries, necessitating a coordinated international response to address shared challenges such as catastrophic risks, bias, and the spread of synthetic content.

Furthermore, the Act encourages cooperation with like-minded international partners. This includes engaging in bilateral and multilateral dialogues, sharing research and development insights, and collaborating on joint initiatives to advance AI safety and security. The goal is to build a global consensus around responsible AI governance, ensuring that while the U.S. maintains its competitive edge in AI innovation, it also contributes to a global ecosystem where AI is developed and deployed safely and ethically. This international dimension is crucial for addressing issues such as cross-border data flows, the development of common evaluation methodologies, and the harmonization of regulatory approaches, ultimately contributing to a more secure and prosperous global digital economy driven by AI.

Implementation Timeline

Compliance Checklist

Sources and References