Trump’s new AI order could change how OpenAI, Google, and Anthropic launch models

In a significant move, President Donald Trump has called on major artificial intelligence firms, including Google, OpenAI, and Anthropic, to voluntarily share early access to their advanced AI models. This initiative aims to facilitate essential cybersecurity evaluations before these technologies are made public. Trump's executive order mandates that various government departments, including Treasury, Defense, Commerce, and Homeland Security, establish cooperative agreements with AI developers. These agreements will focus on safety assessments and cybersecurity tests for AI models. Notably, participation is not compulsory; the voluntary nature of the system allows companies to choose whether to collaborate with federal agencies. The executive order also emphasizes the need for broadened adoption of cutting-edge AI technologies in sectors vital for national security and public services. This expansion will encompass critical infrastructure such as rural healthcare facilities, community financial institutions, and local utilities. Furthermore, the order stipulates that federal agencies may conduct testing on AI models for a maximum of 30 days prior to their public release, a reduction from the previous proposal that allowed for a 90-day review period. Importantly, the order clarifies that it does not create mandatory licensing or preclearance requirements for the development or distribution of new AI models, including those at the forefront of technology. The response from the tech industry has been largely positive. Leaders from companies like Google, OpenAI, and Anthropic have expressed support for the new directive. Microsoft President Brad Smith commented on social media that this executive order represents a crucial advancement in fostering innovation while ensuring public safety. Similarly, Meta’s President, Dina Powell McCormick, noted that the order positions the U.S. at the forefront of AI development, promoting both innovation and security. Additionally, Google’s Kent Walker described the new AI order as a significant advancement, while OpenAI CEO Sam Altman remarked that it effectively finds the right balance in AI development. He added that the U.S. should continue leading in AI by ensuring the development of safe models and equipping trusted defenders with cybersecurity tools.