Blab AI
U.S. AI Policy in Flux: Executive Orders, State Lawsuits, and Export Controls Clash
The United States is currently navigating a fragmented landscape of AI regulation, with federal executive orders, state‑level litigation, and new export controls intersecting in ways that have raised concerns among industry and policymakers.
In early June 2026, Florida’s attorney‑general filed the first state‑level civil action against OpenAI and its chief executive, Sam Altman. The 83‑page complaint, served in Highlands County Circuit Court, accuses the company of prioritizing profit over safety and of marketing ChatGPT in ways that ignored escalating safety warnings. The suit also alleges false advertising and claims that the harms caused by the chatbot are substantial and outweigh any benefits. The lawsuit is part of a broader trend: New York, joined by 41 other states, has issued a sweeping subpoena to OpenAI seeking documents on advertising, user engagement, data handling, and policies related to minors, seniors, and health data.
These state actions come after the federal government’s attempts to shape AI policy have been uneven. President Donald Trump revoked the Biden administration’s Executive Order on AI (EO 14110) in January 2025 and issued EO 14179, titled “Removing Barriers to American Leadership in Artificial Intelligence.” The new order encourages preflight testing of AI systems but makes such testing voluntary and focuses mainly on cybersecurity. Critics argue that the order is too weak to address the broader harms that states are now suing over.
The Florida lawsuit highlights several risks that the Trump order does not address, including safety, transparency, and potential for misuse. The complaint cites incidents such as the 2025 Florida State University shooting, where the perpetrator reportedly used ChatGPT before the attack, and other cases in which chatbots have been linked to violent or self‑harmful outcomes.
In December 2025, Trump signed EO 14365, which directs federal agencies to develop a unified national AI policy, evaluate state laws for conflicts, and condition federal funding on state compliance. The order also exempts state laws related to child safety, data‑center infrastructure, and state procurement.
Meanwhile, the Department of Commerce’s Bureau of Industry and Security issued a new export‑control order on AI products and advanced computing items. The order, announced in January 2025, restricts the export of AI model weights and components of AI software that could be used to train advanced models. The policy is part of a broader effort to limit China’s access to AI technology. The Commerce order also led to the temporary shutdown of Anthropic’s Fable and Mythos models, a move that drew criticism for its abruptness and perceived arbitrariness.
Industry observers note that the export controls are the first to target the numerical inputs of AI models themselves, rather than the hardware used to train them. The policy reflects a shift in U.S. national‑security strategy, but it has also raised concerns about the impact on U.S. AI companies that rely on global supply chains.
The combination of weak federal guidance, aggressive state litigation, and sudden export controls has created uncertainty for AI developers. Some companies argue that voluntary testing and state‑level oversight are insufficient to manage the complex risks associated with large language models. Others contend that the export controls could stifle innovation and disadvantage U.S. firms in a rapidly evolving market.
As the federal government continues to refine its approach, stakeholders are calling for a balanced, bipartisan framework that is technically informed and proactive rather than reactive. The current situation leaves many questions unresolved, including how to enforce safety standards, how to coordinate between federal and state authorities, and how to manage the export of AI technology without hindering domestic innovation.
In the coming months, the U.S. will likely see further developments: additional state subpoenas, possible federal hearings on the export‑control order, and potential amendments to the executive orders that govern AI policy. The outcome of these actions will shape the trajectory of AI development and deployment in the United States.
In early June 2026, Florida’s attorney‑general filed the first state‑level civil action against OpenAI and its chief executive, Sam Altman. The 83‑page complaint, served in Highlands County Circuit Court, accuses the company of prioritizing profit over safety and of marketing ChatGPT in ways that ignored escalating safety warnings. The suit also alleges false advertising and claims that the harms caused by the chatbot are substantial and outweigh any benefits. The lawsuit is part of a broader trend: New York, joined by 41 other states, has issued a sweeping subpoena to OpenAI seeking documents on advertising, user engagement, data handling, and policies related to minors, seniors, and health data.
These state actions come after the federal government’s attempts to shape AI policy have been uneven. President Donald Trump revoked the Biden administration’s Executive Order on AI (EO 14110) in January 2025 and issued EO 14179, titled “Removing Barriers to American Leadership in Artificial Intelligence.” The new order encourages preflight testing of AI systems but makes such testing voluntary and focuses mainly on cybersecurity. Critics argue that the order is too weak to address the broader harms that states are now suing over.
The Florida lawsuit highlights several risks that the Trump order does not address, including safety, transparency, and potential for misuse. The complaint cites incidents such as the 2025 Florida State University shooting, where the perpetrator reportedly used ChatGPT before the attack, and other cases in which chatbots have been linked to violent or self‑harmful outcomes.
In December 2025, Trump signed EO 14365, which directs federal agencies to develop a unified national AI policy, evaluate state laws for conflicts, and condition federal funding on state compliance. The order also exempts state laws related to child safety, data‑center infrastructure, and state procurement.
Meanwhile, the Department of Commerce’s Bureau of Industry and Security issued a new export‑control order on AI products and advanced computing items. The order, announced in January 2025, restricts the export of AI model weights and components of AI software that could be used to train advanced models. The policy is part of a broader effort to limit China’s access to AI technology. The Commerce order also led to the temporary shutdown of Anthropic’s Fable and Mythos models, a move that drew criticism for its abruptness and perceived arbitrariness.
Industry observers note that the export controls are the first to target the numerical inputs of AI models themselves, rather than the hardware used to train them. The policy reflects a shift in U.S. national‑security strategy, but it has also raised concerns about the impact on U.S. AI companies that rely on global supply chains.
The combination of weak federal guidance, aggressive state litigation, and sudden export controls has created uncertainty for AI developers. Some companies argue that voluntary testing and state‑level oversight are insufficient to manage the complex risks associated with large language models. Others contend that the export controls could stifle innovation and disadvantage U.S. firms in a rapidly evolving market.
As the federal government continues to refine its approach, stakeholders are calling for a balanced, bipartisan framework that is technically informed and proactive rather than reactive. The current situation leaves many questions unresolved, including how to enforce safety standards, how to coordinate between federal and state authorities, and how to manage the export of AI technology without hindering domestic innovation.
In the coming months, the U.S. will likely see further developments: additional state subpoenas, possible federal hearings on the export‑control order, and potential amendments to the executive orders that govern AI policy. The outcome of these actions will shape the trajectory of AI development and deployment in the United States.
