Microsoft President: AI Needs Global Safety Rules, Not Self-Regulation

Overview

Microsoft President Brad Smith has issued a stark warning regarding the current state of artificial intelligence regulation, stating that everyday consumer goods like milk and washing machines are subject to more rigorous safety oversight than modern AI systems. He argued that the tech industry must move beyond voluntary self-regulation and adopt binding global safety rules to mitigate emerging risks.

Smith’s comments place him in direct contrast with President Donald Trump, who has advocated for minimal government interference to ensure American companies remain competitive against Chinese rivals. The Microsoft executive emphasized that the window for proactive safety measures is closing, calling it "the moment" for the world to establish firm guardrails for powerful AI technologies.

The Case for Global Regulation

Smith suggested that existing frameworks for consumer product safety should serve as the blueprint for AI governance. He pointed out that every carton of milk sold in grocery stores and every washing machine in households undergoes strict safety inspections, whereas AI developers currently operate with far fewer mandatory checks.

"Right now, your washing machine is subject to more safety rules than AI. Every carton of milk that you see in the grocery store is subject to safety rules and inspection."

The president of Microsoft outlined several specific regulatory requirements that he believes should be implemented globally. These include a legal obligation for companies to notify victims if a rogue AI agent breaches their systems. Additionally, he called for the appointment of independent government evaluators to vet new AI products before they are released to the public.

Smith also proposed that AI systems require "emergency brakes," commonly referred to as kill switches, which allow developers or operators to pause or slow down a system if it begins to behave unpredictably or dangerously. He argued that incident reports should be mandatory, creating a transparent record of when AI systems fail or cause harm.

Critics, including those in the current White House, worry that heavy-handed regulations could stifle innovation. However, Smith countered that these measures are essential for long-term stability and public trust. He noted that the leaders of major US-based AI firms, including Anthropic, OpenAI, and Google DeepMind, have previously backed calls for an AI development slowdown, signaling a growing consensus within the industry about the need for caution.

Tension with White House Policy

Despite the internal consensus among many AI leaders, Smith’s position highlights a significant rift with President Trump’s approach to technology policy. Trump has repeatedly stated that whoever leads in AI wins, pushing for the United States to go "full-steam ahead" without the constraints of strict regulatory frameworks. During a recent meeting at the White House, Trump hosted executives from OpenAI, Anthropic, Nvidia, SpaceX, Meta, and Google, who signed a "morally binding" document committing to work with independent auditors.

Smith acknowledged the importance of this voluntary agreement but stopped short of endorsing it as a complete solution. He applauded the statement and confirmed that Microsoft is happy to sign it, describing it as a necessary step. However, he cautioned that self-regulatory regimes often only succeed in persuading responsible companies to comply, while leaving irresponsible actors unchecked.

The president’s view is that external regulation could cause US firms to lose their edge to Chinese competitors. Smith, however, expressed hope that the US government might be more receptive to international standards originating from allies rather than adversaries. He suggested that Washington views the United Kingdom quite differently from how it views Brussels, noting that the UK is seen as a partner rather than a regulatory overreacher.

Britain as a Potential Regulatory Leader

In Smith’s assessment, the United Kingdom is best positioned to drive the creation of global AI safety rules. He cited the UK’s AI Security Institute, launched during the tenure of former Prime Minister Rishi Sunak, as a prime example of technical leadership in the field. When asked to name the country with the strongest technical capability in this area, Smith indicated that most people in the tech sector would immediately point to the United Kingdom.

He argued that the UK could play a pivotal role in bringing different stakeholders together to establish these standards. Smith encouraged a calm but firm approach to diplomacy, hoping that Microsoft and the British government could help bridge the gap between rapid technological advancement and necessary safety precautions.

The discussion around AI safety has intensified in recent months after experts warned that the technology poses a potential threat to humanity. High-profile figures in the sector have raised alarms about the possibility of AI agents escaping human control or models becoming so powerful that they jeopardize global cybersecurity.

Smith’s remarks underscore the urgency felt by many within the technology industry. While the path forward remains contentious, with commercial interests often clashing with safety concerns, the call for a structured, international regulatory framework is gaining momentum. As AI continues to integrate deeper into critical infrastructure, the debate over how to manage its risks will likely define the next era of tech policy.