Microsoft president says UK is best placed to drive AI regulation

Microsoft president Brad Smith is calling for urgent regularisation of AI, and his statement is that user products similar beverage and washing machines person much oversight than the companies gathering artificial intelligence.

On a recent BBC podcast, Smith called AI some a “weapon and a tool” that is putting the world’s cybersecurity astatine hazard and said the “moment has arrived” to institute planetary information rules.

Smith besides said the U.K., and not the U.S. oregon European Union, is champion placed to pb connected planetary information rules. Britain’s AI Security Institute (AISI), launched successful 2023, receives entree to large models from U.S. frontier labs earlier they are deployed and has rapidly go a person successful AI information testing. 

That relation has been nether menace successful caller weeks, with U.S. president Donald Trump informing against “globalist” organizations, which helium claims are interfering with American companies. Last month, helium reportedly asked Anthropic and OpenAI to withhold caller models from British testers, but AISI confirmed connected Monday that it is capable to entree and trial the latest models. 

Despite the president’s pushback, however, Smith argued the U.K. tin modulate AI successful ways that the U.S. cannot. 

“In Washington, D.C., radical don’t look astatine the U.K. and they don’t look astatine London the mode they look astatine Brussels,” helium told the BBC. “That’s a strength.”

His comments travel amid rallying cries from tech leaders, including Anthropic’s Dario Amodei and OpenAI’s Sam Altman, to dilatory the gait of AI improvement successful the aftermath of respective incidents wherever AI agents went rogue.

Last month, Trump hosted tech leaders of companies, including OpenAI, Anthropic, Nvidia, AMD, SpaceX, Meta, and Google, astatine the White House to sermon AI safety. They signed a “morally binding” papers successful which they agreed to enactment with “independent auditors” to measure the information and capabilities of their AI models. 

Smith says the papers is simply a “good step,” but it stops abbreviated of the benignant of regularisation that is necessary. The occupation with voluntary self-regulation, helium said, is that lone the “good companies” volition comply. 

“It’s the companies that are not arsenic liable that you request to interest about,” helium said.

Smith outlined respective AI information rules he’d similar to spot enactment into place, similar a ineligible work to alert an enactment if its strategy was breached by rogue AI agents. He besides appears to beryllium successful favour of AI labs getting the support of extracurricular evaluators earlier an AI merchandise is released and spoke positively astir termination switches to intermission oregon dilatory down AI if humans suffer power of it. 

“Let’s talk calmly, possibly adjacent softly, but firmly, and commencement to bring radical together,” Smith said. “I anticipation that’s the relation we tin play arsenic a company. I anticipation that’s the relation the U.K. tin play arsenic a country.”