AI Policy Newsletter - May 5, 2026

• News broke on Monday that the Trump administration is considering issuing a new to create an AI working group to develop possible government oversight of new AI models including certain safety standards. The development comes in the wake of Anthropic’s limited release of its new Claude Mythos model, which has evoked concerns around the exploitability of powerful AI models. On May 5, Google, Microsoft, and xAI to provide the federal government with access to new AI models before they are deployed in order to conduct tests and evaluations of the capabilities of the models.

• On April 30, the Senate Judiciary Committee the GUARD Act, legislation that would impose criminal penalties on AI companies whose chatbots engage in sexually explicit content with or encourage self-harm or violence from minors and ban AI companies from providing emotional AI chatbot companions for children. House companion legislation is being by Reps. Blake Moore (R-UT) and Valerie Foushee (D-NC). Sen. Ted Cruz (R-TX) voted yes but broke with the bill’s outright ban on minors using chatbots, arguing some applications are beneficial – and separately the CHATBOT Act with Sen. Brian Schatz (D-HI), which takes a narrower approach by requiring parental family account controls for users under 13.

• On April 23, Director of the Office of Science Technology and Policy Michael Kratsios, released a memo titled, “ In the memo, Kratsios outlined the Trump administration’s strategy for addressing threats from adversaries, particularly China, from illegally distilling U.S. innovations to create their own models and strip out security protocols. The strategy relies on public-private sector information sharing and collaboration to create strong defenses against attacks and explore ways to keep foreign adversaries accountable.

Download the full newsletter