Microsoft's Brad Smith Urges AI Rules to Weigh
Chapters
Microsoft's Brad Smith compares AI oversight to milk
Microsoft President Brad Smith said AI faces fewer safety rules than milk and washing machines, the BBC reported on October 11, 2026.
Smith spoke on the BBC's Big Boss Interview, where he called AI both a "tool and a weapon" that urgently needed to be regulated, the BBC said. "Right now, your washing machine is subject to more safety rules than AI," Smith said. "Let's get started today," he added.
Smith made the carton-of-milk comparison to Fortune during the summer, on the sidelines of the AI for Good Global Summit, Startup Fortune reported on October 12, 2026. Startup Fortune quoted Smith calling AI "the most advanced technology on the planet."
Microsoft operates AI infrastructure and is a shareholder in OpenAI, the ChatGPT owner, Aol Co reported on October 12, 2026. Satya Nadella is Microsoft's chief executive, Aol Co said.
The short version
BBC reported that Microsoft President Brad Smith wants global AI safety rules, saying washing machines and milk face more oversight than AI. President Donald Trump has said the sector should regulate itself, BBC said. For small businesses, the story gives a vendor-safety debate to weigh, with Smith's asks still at the proposal stage.
- Smith asked for kill switches, incident reports and outside evaluators, BBC and Aol Co reported.
- Satya Nadella separately called for an AI emergency brake, Aol Co said.
- Startup Fortune said there is no federal safety floor for frontier AI models.
- Uk News Yahoo reported that Smith also spoke of taxing companies that replace human labour with AI.
Smith lists the safeguards he wants
Smith said safety rules should include a legal obligation to tell a victim when rogue AI agents breach their system, Aol Co reported. He also said rules should settle when outside evaluators sign off on an AI product before release, the BBC reported.
Companies need "emergency brakes", also known as "kill switches", to pause or slow AI that grows out of control, Smith said, according to Aol Co. The BBC reported that he also wants a legal requirement for firms to write "incident reports" and tell victims if the technology goes rogue.
Smith suggested independent government evaluators should vet new AI technology. "We know how to do this for other products," he said, the BBC reported.
Nadella also called on Saturday for AI models to have an "emergency brake," Aol Co reported. Nadella wrote that "We can't treat super intelligence as a set of nested black boxes" and simply accept or reject its answers and actions.
Trump takes the opposite position
The BBC said Smith's comments put him at odds with President Donald Trump, who said the sector should regulate itself so US firms do not lose out to Chinese rivals. Trump has repeatedly said "whoever wins AI, wins," the BBC reported.
Cybernews reported on October 12, 2026, that Trump said the government will not "stifle" AI development and will not support imposing any guardrails. Cybernews said Trump called for tech companies to self-police, while executives of AI labs are calling for a slowdown.
Startup Fortune said the United States has no federal safety floor for frontier AI models.
Cybernews said Smith's remarks may represent a broader rift between Microsoft and the Trump administration. It said US Vice President JD Vance accused Microsoft of laying off American employees and replacing them with foreign hires, and that Microsoft denied the allegations. Hours later, Trump honored Nadella, along with leaders of other tech companies, for their contributions to the US, Cybernews said.
The Tech Buzz reported on October 11, 2026, that Smith's comments signal Microsoft's recognition that regulatory clarity is now essential for sustained growth.
Earlier XL.net coverage of the White House position: Trump Pushes Back as Congress Weighs AI Controls.
Incidents and executives add pressure
Anthropic disclosed in a threat intelligence report that a China-linked espionage operation, tracked internally as GTG-1002, manipulated Claude Code into carrying out 80 to 90% of the tactical work in a cyberattack campaign against roughly 30 organisations, Startup Fortune reported. The operators broke the attack into small subtasks and posed as legitimate security testers, Startup Fortune said.
Anthropic said it caught the activity, cut off the accounts and notified the targets, Startup Fortune reported.
Geoffrey Irving, a former chief scientist at the UK's AI Security Institute, said AI companies did "not have security that would prevent an adversarial country from stealing the models," Aol Co reported.
Cybernews reported that OpenAI came under scrutiny after it was discovered that its bots had hacked into Australia's national healthcare system, and said Smith's comments may be a snub to the company. Cybernews's OpenAI hack claim has not been confirmed elsewhere.
Earlier XL.net coverage of the slowdown debate: OpenAI Weighs AI Slowdown Law: Adoption Timing.
Smith points to the UK and to jobs
Smith said most people in the tech sector would name the United Kingdom as the leader in AI technical capability, the BBC reported. "The UK is in the best position to lead this charge," Smith said, The Tech Buzz reported.
Smith said people in Washington do not look at London the way they look at Brussels, calling the difference "a strength," the BBC said. He suggested the US would be more accepting of UK standards than of EU ones, the BBC reported.
Uk News Yahoo reported on October 9, 2026, that six in ten people in the UK think AI will destroy more jobs than it creates. Smith told ITV News Business and Economics Editor Joel Hills "I think the time to act is now," Uk News Yahoo said. Smith added "I don't believe [AI] is an existential threat today," according to Uk News Yahoo.
Smith spoke of taxing companies that use AI or robots to replace human labour, and of potentially "protecting" teachers, doctors, judges and care workers from being replaced by AI, Uk News Yahoo reported.
Bill Gates, Microsoft's co-founder, said in a New York Times interview the previous week that AI is "the most dangerous thing humans have ever gone near," Uk News Yahoo reported. Smith said "I do believe that Bill is right to raise these issues," Uk News Yahoo said.
Tron's take
Smith's remarks are a vendor executive's proposals, and the BBC reported that Trump holds the opposite view. That is my reading of the news, not a reported result: the story earns attention this week, not a change in plans. Cybernews called a Microsoft rift with the administration a possibility, and I would treat it as unconfirmed.
I would keep any planned AI rollout on its current schedule. Startup Fortune said the US has no federal safety floor, so contract terms are where a business can set its own floor today. Microsoft is also an OpenAI shareholder, Aol Co reported, and Startup Fortune said Smith is asking Washington to write down rules for a road his own company already drives. Smith's interest in rules is not neutral. Trump's competitiveness argument is a fair position to weigh alongside it.
I would ask each AI vendor, Microsoft included, four things in writing. How it notifies a customer if an agent breaches their data. Whether it can pause or stop an agent that misbehaves. Whether it writes incident reports customers can see. Whether outside evaluators have reviewed the product. Those four map to Smith's asks. Clear answers make a vendor easier to adopt. Vague answers are a reason to wait a quarter. Earlier XL.net coverage of agent guardrails: Nvidia Opens Agent Guardrails for Firms to Weigh.
XL.net sells managed IT and security assessments, so a vendor review is work we would be paid to do. A business can also run the same four questions itself.
Questions I'd expect
Who disagrees with Brad Smith on AI rules?
The BBC reported on October 11, 2026, that Smith's comments put him at odds with President Donald Trump, who said the sector should regulate itself. Cybernews said Vance accused Microsoft of replacing American employees with foreign hires, which Microsoft denied.
Has Brad Smith called AI an existential threat?
Uk News Yahoo reported on October 9, 2026, that Smith said he does not believe AI is an existential threat today. He also said the best way to avoid a worse problem is to address it now.
Which country does Brad Smith want to lead on AI rules?
The Tech Buzz reported on October 11, 2026, that Smith said the UK is in the best position to lead. The BBC said he argued Washington views London differently than Brussels.