Microsoft's Brad Smith: The US is Regulating AI Without Clear Rules
Microsoft president Brad Smith has expressed concern about how Washington regulates artificial intelligence (AI) without a clear rulebook. "Without rules, businesses can't plan," Smith told Fortune.
Key Takeaways
- Unclear Regulations: Smith argues that the US government is implementing AI regulations without transparent or complete guidelines, hindering business planning.
- Recent Actions: His comments follow two significant moves by the Trump administration:
- The Commerce Department used export control laws to restrict Anthropic's Fable 5 and Mythos 5 models worldwide, citing cybersecurity risks.
- Officials pressured OpenAI to delay the public launch of its GPT-5.6 family, initially making it available only to government-vetted partners.
- Tool Misuse: While Smith acknowledges the need for action on cybersecurity concerns, he criticizes the government's use of a single regulatory tool (export controls) that may not be suitable for modern AI models.
- Lack of Legislation: Legal experts note that export control laws predate API-served models, raising doubts about their applicability in court.
- Global Impact: The Anthropic incident sparked discussions about sovereign AI, with governments considering controlling AI models and infrastructure to ensure supply certainty.
- Shared Responsibility: Smith suggests that US firms and Washington must now prove the reliability of their AI systems to global users.
Diverging Perspectives
Salesforce chief Marc Benioff took a different view at the same event, feeling "good" about the export controls and suggesting Europe had misconstrued a security concern as hostility.