Can UK Mandate AI Safety Standards After Microsoft President Highlights Regulatory Gaps?
Brad Smith argues AI faces fewer regulations than dairy products, urging the UK to lead in establishing mandatory global safety standards.
Key points
- Microsoft president Brad Smith states that milk and washing machines are subject to more safety rules than artificial intelligence systems.
- He calls for mandatory incident reporting, independent model evaluations, and emergency kill switches for rogue AI agents.
- Smith believes the UK is well-positioned to lead on AI regulation due to the lack of strict US government action.
Microsoft president Brad Smith has stated that artificial intelligence systems currently face fewer safety regulations than everyday consumer goods like milk and washing machines. In interviews with ITV and the BBC, Smith emphasized that while dairy products undergo strict inspection, AI lacks comparable mandatory oversight. He argued that this regulatory gap poses significant risks to global cybersecurity and requires immediate attention from policymakers.
Why does Smith compare AI regulation to milk and appliances?
Smith used the comparison to highlight the disparity in safety standards between high-risk technology and low-risk household items. He told ITV that every carton of milk in a grocery store is subject to safety rules and inspection. He added that washing machines are also governed by more safety protocols than AI systems. This analogy underscores his concern that powerful AI models, which could potentially hack companies and governments, operate without equivalent legal safeguards.
What specific safety measures is Smith proposing?
Smith called for stronger global safety rules that would legally require AI companies to notify victims when their systems are breached by rogue agents. He also advocated for mandatory reporting of safety incidents and the implementation of independent evaluators to review new models before release. Additionally, he suggested the need for "emergency brakes" or kill switches to stop systems that go rogue. Microsoft CEO Satya Nadella echoed the call for emergency brakes, stating humans should be able to pause or shut down AI models mid-task.
How does the US approach differ from Smith's proposal?
The US approach, supported by President Donald Trump, favors industry self-regulation over government mandates. Trump has argued that excessive regulation could disadvantage the US in the AI race against China and has urged the industry to accelerate development. At a White House meeting on September 29, executives from major tech firms signed a voluntary agreement outlining limited safety standards. Trump described these standards as "morally binding," though critics noted the accord lacks legal enforcement mechanisms and allows firms to define safety themselves.
Why might the UK lead on AI regulation?
Smith believes the lack of decisive US action creates a unique opportunity for the UK to lead in AI regulation. He noted that Washington does not view London with the same skepticism it holds for Brussels, which he described as a strength for the UK. Smith expressed hope that the UK could play a leading role in establishing global standards, particularly as AI models become powerful enough to threaten world cybersecurity. He stated that the moment for such regulatory action has arrived.
What to do and how to stay safe: Microsoft
- Monitor official channels from your organization's security team for updates on AI safety policies and incident reporting procedures.
- Ensure that any AI tools integrated into your development workflow have clear kill switches or pause mechanisms for immediate human intervention.
- Review vendor contracts to understand liability and notification protocols in the event of an AI-related security breach or rogue behavior.
- Stay informed about emerging regulatory frameworks in the UK and US that may impact compliance requirements for AI deployment.
Step-by-step guide: Build an Incident Response Plan That Actually Works
General security guidance from the Sandbox Dispatch newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Who is calling for stricter AI safety regulations?
Microsoft president Brad Smith is calling for stronger global safety rules, including mandatory incident reporting and independent model evaluations.
What did tech executives sign at the White House?
Executives from Google, Meta, Anthropic, OpenAI, xAI, and Nvidia signed a voluntary agreement outlining limited AI safety standards on September 29.
Does the US government plan to regulate AI strictly?
No, President Donald Trump has advocated for industry self-regulation, arguing that strict rules could hinder the US in the AI race against China.
