The Trump administration has issued an executive order urging AI developers to voluntarily submit their advanced models for cybersecurity testing prior to public release. This move, prompted by escalating security concerns over powerful AI systems like Anthropic’s Mythos, directs key government departments—including Treasury, Defense, Commerce, and Homeland Security—to establish agreements with AI firms. Agencies will have a maximum of 30 days to conduct tests on these models, which is part of a broader initiative to enhance cybersecurity across federal operations. The order marks a notable shift in Trump's approach to AI oversight, transitioning from a hands-off policy to a more proactive stance on monitoring AI capabilities. However, this voluntary testing could potentially hinder the industry's profits by delaying new model launches or necessitating changes to address security vulnerabilities. Notably, companies such as OpenAI, Anthropic, and Google have previously engaged with the government regarding cybersecurity measures.