LONDON, Oct 12 (Reuters) – Artificial Intelligence needs urgent global regulation with independent evaluators able to vet new developments, the president of Microsoft said, warning that AI currently faces fewer safety requirements than milk.
In an interview with the BBC published on Monday, Brad Smith said a set of global safety rules needed to be in place, requiring a “victim” to be informed if their system is breached by a rogue AI agent and an “emergency brake” available.
He said there should also be a legal requirement for companies to write “incident reports”.
“We know how to do this for other products,” Smith said. “We should build on that knowledge and think about how to apply it to AI.”
“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.”
Tech companies such as Anthropic and OpenAI have started disclosing rogue or undesired behaviour by AI agents as part of a global debate about the fast-advancing technology that has included researchers’ warnings of an eventual existential threat to humanity.
Some companies have called for AI development to slow down, but that has sparked a debate as to whether executives have ulterior motives such as setting rules that preserve the lead of their firms, pre-empting a broader government backlash, or simply slowing breakneck spending.
Microsoft’s Smith said the one country that was well-placed to lead the way on regulation of AI was Britain, due to the AI Security Institute which was launched under former prime minister, Rishi Sunak.
Smith said officials in Washington viewed London very differently to how they viewed Brussels, which regulates AI based on the level of risk via its AI Act.
Britain’s institute receives pre-deployment access under voluntary agreements with the major AI models, allowing it to assess their capabilities and risks.
AI Minister Kanishka Narayan has told Reuters that Britain would consider regulating advanced models if the current system proved insufficient.
(Reporting by Kate Holton;Editing by Alison Williams)
Comments