Monday 12 October 2026
Evening Edition
ZOTPAPER

News without the noise

ZOTPAPER
AI

Microsoft president warns AI is less regulated than milk and urges urgent safety rules

Brad Smith backs 'kill switches' and liability for harm as industry leaders urge slower development

By LineZotpaper
Published
Read Time2 min
Microsoft president Brad Smith has warned that artificial intelligence is subject to fewer safety rules than milk or washing machines, calling for urgent regulation including 'emergency brakes' and legal liability for harm caused by rogue AI agents.

Brad Smith told the BBC that AI needed more safety rules, including a legal obligation to a 'victim' if their system was breached by rogue AI agents.

'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,' Smith said.

Smith described AI as a potential 'tool and a weapon' and said companies needed 'emergency brakes', also known as 'kill switches', to pause or slow AI if it grew out of control.

His comments came alongside those of Microsoft chief executive Satya Nadella, who on Saturday called for AI models to have an 'emergency brake'. Nadella wrote that superintelligence could not be treated as 'a set of nested black boxes' and that developers must build 'contained systems whose behaviour we can observe, limits we can test, and actions we can always contain'.

Microsoft, which operates AI infrastructure and is a shareholder in OpenAI, is led by Nadella.

The warnings follow growing calls from experts, politicians and regulators for tougher controls on AI, and come after Anthropic, Google DeepMind and OpenAI cautioned that the pace of the technology's development should slow. Britain's AI minister, Kanishka Narayan, was also preparing to call on firms developing the most advanced systems.

§

Analysis

Why This Matters

  • A senior executive at one of the world's largest AI companies is publicly arguing that the technology is dangerously under-regulated, a notable shift from within the industry.
  • The comparison to everyday consumer goods frames the debate around enforceable safety standards and liability for harm.
  • The call for emergency brakes targets the next phase of AI: autonomous agents that can act without direct human oversight.

Background

AI models have grown rapidly more capable, and companies are now deploying 'agents' that can carry out tasks independently. Governments are still working out how, or whether, to regulate the technology. Microsoft is an unusual voice in this debate because it both builds and operates AI systems and holds a stake in OpenAI, the company behind ChatGPT. Its leadership is now arguing that the industry cannot rely on self-regulation alone.

Key Perspectives

Microsoft: Smith says AI needs safety rules comparable to those for milk and washing machines, including a legal obligation to victims if a system is breached by rogue agents. Nadella argues for 'contained systems' whose actions can always be checked or stopped. Anthropic, Google DeepMind and OpenAI: These developers have warned that the pace of AI development should slow, adding weight to safety concerns. Governments and regulators: Experts, politicians and regulators are increasingly calling for tougher rules. Britain's AI minister, Kanishka Narayan, was preparing to press firms on safety as the warnings were made.

What to Watch

  • Whether Britain or other governments turn these calls into binding safety requirements.
  • How quickly regulators define liability for harm caused by autonomous AI agents.
  • Any concrete adoption of 'kill switch' or emergency brake mandates in new AI legislation.

Sources

Zotpaper

Written by software from the reporting listed above, scored by an automated standards desk, and published without a person reading it first. If something here is wrong, tell the editor and it will be put right.

How we workSubscribe