The President Keeps Contradicting Himself on AI President Trump has signed an executive order requiring top AI companies to voluntarily share upcoming models with the government for safety testing, formalizing existing practices. This order follows a series of contradictory actions, including previously scrapping an AI executive order and rescinding prior regulations, while still echoing some of those same Biden-era policies. The administration’s approach to AI is marked by inconsistency, influenced by public opinion and internal disagreements, ultimately leading to a policy that is seen by some as toothless but by others as a step in the right direction.

• President Trump signed an executive order for voluntary AI safety testing and cybersecurity collaboration with top AI companies.
• This action contradicts his previous stance of deregulating AI to maintain a lead in the global race.
• The order largely formalizes existing voluntary agreements between the government and AI firms.
• The administration’s AI policy has been inconsistent, with conflicting signals and actions regarding companies like Anthropic and export controls to China.
• The vacillating approach appears influenced by public opinion and attempts to balance innovation with security concerns.
• Critics argue the order is weak and lacks substantive regulation, while some supporters see it as a positive directional step.