Microsoft president Brad Smith has criticized the US approach to AI regulation, highlighting the absence of clear and visible guidelines. He warns that this regulatory ambiguity hampers business planning and poses challenges for international confidence in AI technology.
- US uses outdated export controls to regulate AI without transparent rules
- Regulatory uncertainty hinders business planning and global trust
- Government’s list of trusted AI partners and future oversight remains unclear
What happened
Brad Smith, president of Microsoft, has publicly highlighted the current state of US AI regulation as lacking clear, transparent, and comprehensive rules. He made his comments at the AI for Good Global Summit, emphasizing that businesses cannot effectively plan or operate in such an uncertain regulatory environment. Smith pointed to recent US government interventions as examples of this challenge.
Last month, the US Commerce Department applied export control laws to restrict distribution of Anthropic’s AI models, citing cybersecurity concerns. Similarly, officials delayed the public release of OpenAI’s GPT-5.6 models, allowing early access only to vetted government partners. While some restrictions have eased, these actions exposed the administration’s reliance on a regulatory tool that predates contemporary AI technologies.
Why it matters
Smith acknowledges that addressing cybersecurity risks promptly is necessary, but he cautions that the government’s exclusive use of export controls is a blunt instrument ill-suited for nuanced AI oversight. Legal experts question whether such controls can hold up under legal scrutiny since they did not anticipate API-driven AI models when established. This creates a regulatory regime that lacks legislative support and transparency.
This opacity fuels uncertainty within the AI sector, making it difficult for companies to navigate compliance and invest confidently. Moreover, it complicates international dynamics, as sovereign AI initiatives react to US actions perceived as restricting foreign access. Smith stresses that US firms and policymakers must build confidence in their AI systems to maintain global supply assurance and market trust.
What to watch next
Internationally, responses to US regulatory actions may accelerate sovereign AI efforts as countries pursue independent controls over their AI infrastructures. How Washington and American companies address these trust and supply concerns will shape the future global AI landscape, influencing policy development, market access, and international cooperation.