Microsoft President Brad Smith has called for urgent global regulation of artificial intelligence, warning that autonomous AI systems face fewer statutory safety mandates than everyday consumer goods like washing machines and milk. In an interview with the BBC, Smith urged governments to implement standardized AI safety frameworks incorporating mandatory incident reporting, mandatory disclosure to users affected by security breaches, independent third-party model evaluations, and automated kill-switch mechanisms for rogue AI agents.
Highlighting international policy approaches, Smith positioned the United Kingdom as uniquely suited to spearhead global AI standards via its AI Security Institute—which holds pre-release testing agreements with major foundation model developers. He noted that US officials view the UK’s collaborative, safety-tested framework more favorably than the European Union’s rigid risk-classified AI Act. Smith’s comments follow recent reports from OpenAI and Anthropic regarding unexpected behaviors in autonomous AI agents, alongside emerging industry testing around next-generation models like Google’s Gemini 4 Carbon.
Smith’s advocacy arrives amid intense industry debate over whether stringent regulatory requirements genuinely promote public safety or serve to insulate incumbent tech giants by raising compliance costs for open-source rivals. While UK AI Minister Kanishka Narayan confirmed the government remains open to introducing binding legislation if voluntary industry commitments prove insufficient, technology companies and policymakers continue to grapple with balancing autonomous AI guardrails against global competitiveness.
Leave a Reply