Microsoft President calls for tighter AI regulation, saying safety regulations are worse than cartons of milk

Brad Smith, president of global technology giant Microsoft (Microsoft), said that artificial intelligence (AI) urgently needs global supervision, and an independent agency should be established to review the latest developments in AI technology. He warned that there are currently even fewer safety regulations related to AI than milk.

In an exclusive interview published by the British Broadcasting Corporation (BBC) on the 12th, Smith said that the world should develop a set of security regulations that require that when a system is invaded by a malicious AI agent, the affected people must be notified and an "emergency braking" mechanism must be set up.

He also said the law should require companies to submit "incident reports."

"We know how to do this with other products," Smith said. "We should learn from that experience and think about how to apply these practices to AI."

He pointed out: "Currently, washing machines are subject to more safety regulations than AI. Even every carton of milk you see in the store is subject to safety regulations and inspections."

Technology companies including Anthropic and OpenAI have begun to expose anomalous or unexpected behavior of AI agents. The cases are part of a global debate over rapidly developing AI technology, which has also included warnings from researchers that AI could threaten humanity's survival.

Some companies have called for slowing down the development of AI, but this has also raised questions about whether corporate executives have other considerations, such as formulating rules that will help maintain their company's leading position, preempting broader government backlash, or simply hoping to slow down too urgent spending.

Smith pointed out that the UK is one of the countries best positioned to lead AI regulation because of the AI ​​Security Institute established during the tenure of former Prime Minister Rishi Sunak.

Smith pointed out that Washington officials view London very differently from Brussels.

The European Union (EU) regulates AI according to the degree of risk according to the Artificial Intelligence Act (AI Act).

The British AI Security Research Institute, through voluntary agreements with major AI model developers, obtains access rights before the model is officially deployed to assess its capabilities and risks.

British AI Minister Kanishka Narayan previously stated in an interview with Reuters that the UK will consider regulating advanced AI models if the current system proves insufficient to cope with related risks.