On June 3, 2026, President Donald Trump signed an executive order aimed at establishing a process for leading AI companies to voluntarily share upcoming models with the government for safety testing prior to their wider release. The order follows an earlier plan that was canceled shortly before a signing ceremony, where Trump emphasized the importance of maintaining the U.S. leadership in AI. The new policy requests collaboration from companies like OpenAI and Anthropic to enhance cybersecurity measures.
Experts, including Daniel Remler from the Center for a New American Security, noted that the order formalizes existing agreements between the government and AI firms regarding safety testing. While the order has been described as less stringent than earlier proposals, it reflects a shift in the administration's approach to AI regulation.
The executive order has drawn mixed reactions, with some praising it as a step forward while others criticize it for lacking enforceable measures. The administration's inconsistent stance on AI has been highlighted, particularly regarding its relationship with companies like Anthropic and its regulatory approach towards China. As public concern over AI grows, the administration's ability to implement effective regulation may be challenged by the rapidly evolving landscape of AI technology.
