ProductIndependently confirmed2 publishers2 min readPublished Updated
Microsoft president Brad Smith wants a legal duty to tell victims when a rogue AI agent breaches them
Microsoft's Brad Smith told the BBC that milk and washing machines face more safety rules than AI. The BBC says his call for a legal duty to tell breach victims, and for outside sign-off, puts him at odds with Trump, who says the sector should regulate itself.
The Product Desk

What happened
- Smith said firms need "emergency brakes", which the industry calls kill switches, to pause or slow an AI system that gets out of control.
- At the end of last month, leaders of OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google signed a "morally binding" White House document committing them to work with independent auditors.
- Smith said he applauds that statement and that Microsoft is "happy to sign it".
- He said the UK is best placed to lead global rules because of its AI Security Institute, which began as the AI Safety Institute under Rishi Sunak.
- Smith suggested the US would accept UK-made standards more readily than the EU's, saying Washington does not look at London the way it looks at Brussels.
Why it matters
- contradiction The gap with Trump is narrower than "at odds" suggests: the White House document and Smith both back independent auditors, and Smith's objection is that a voluntary pledge is necessary but not sufficient.
- decision Teams running agents have to choose between building a pause control and an incident record now, on their own terms, or waiting for a law that sets the terms.
- precedent Smith's template is existing product safety, so any duties that follow from his proposals would look like the inspection and incident reporting other products already carry.
"Right now, your washing machine is subject to more safety rules than AI," Smith said. "Every carton of milk that you see in the grocery store is subject to safety rules and inspection." [2] He was speaking on the BBC's Big Boss Interview, where he called AI both a "tool and a weapon". [9]
Each item on his list turns into work for someone. A pause control needs an owner who has tried it, and a notice to a victim needs a record of what the agent did and whom it touched. Smith described the brake and the incident report as things firms should be required to have. [4][5]
The document signed at the White House and the rule Smith wants are different instruments. The first is described as "morally binding"; the second is a legal requirement. [11][5] Trump's stated reason for leaving it to industry is competition: he said the sector should regulate itself so US firms do not lose out to Chinese rivals. [10]
Smith's objection to the voluntary route is about who it reaches. He said the challenge for any "self-regulatory regime" is that "the good companies will do what you want". [14] He added: "It's the companies that are not as responsible that you need to worry about." [15]
Smith framed outside sign-off as a question of when it should apply. [3] The BBC report does not include his answer, or a definition of a rogue-agent breach.
We'd build the brake and the incident record before any law asks for them, and treat outside sign-off as a question to put to vendors. The tradeoff is time before launch, since each needs an owner and a test. A two-by-two sorts the agents you already run. One axis is whether someone on your team can pause the agent in a step that has actually been tried. The other is whether you can produce a record of what it did and whom it touched.
An agent with both can keep the access it has. An agent with a tested pause but no record should stay away from customer data, because you could stop it and still could not tell a victim what happened. An agent with a record but no pause should be read-only until a brake exists. An agent with neither is not ready for production.
For vendors, add Smith's third item as a purchasing question: whether anyone outside the company has evaluated the product, and what that evaluator signed off.
What to watch
- Whether Microsoft signs the White House document that Smith said it is "happy to sign".
- Any draft text that defines a rogue-agent breach, sets a deadline for telling victims, or says which products need outside sign-off.
- Whether Washington responds to Smith's suggestion that the UK, through its AI Security Institute, lead global AI rules.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence45
- Adoption
- Insufficient
- Hype gap+15
- Incentives
- Insufficient
- Confidence60
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
Brad Smith, president of Microsoft, said milk and washing machines are currently more regulated than AI.
ReportedSupportedSource: BBC, reporting Smith's interview2 sources— create a free account to open themView cited source - [2]
Smith said: "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."
ReportedSupportedSource: Brad Smith, Microsoft president, to the BBC2 sources— create a free account to open themView cited source - [3]
Smith said there needed to be global safety rules, including a legal obligation to tell a victim if their system is breached by a rogue AI agent and a decision on when outside evaluators should sign off on an AI product before release.
ReportedSupportedSource: BBC, reporting Smith2 sources— create a free account to open themView cited source - [4]
Smith said firms need "emergency brakes", often known in the industry as kill switches, to pause or slow down AI if it got out of control.
ReportedSupportedSource: BBC, reporting Smith2 sources— create a free account to open themView cited source - [5]
Smith said there should be a legal requirement for firms to write incident reports and tell victims if the technology goes rogue, and suggested independent government evaluators should vet new AI technology.
ReportedSupportedSource: BBC, reporting Smith2 sources— create a free account to open themView cited source - [6]
Smith said the UK is best placed to drive forward global AI rules because of its AI Security Institute, launched under former prime minister Rishi Sunak's government as the AI Safety Institute.
ReportedSupportedSource: BBC, reporting Smith2 sources— create a free account to open themView cited source - [7]
Smith suggested the US would be more accepting of the UK setting AI standards than of the EU doing so, saying people in Washington DC do not look at London the way they look at Brussels.
ReportedSupportedSource: BBC, reporting Smith2 sources— create a free account to open themView cited source - [8]
Smith said: "We know how to do this for other products. We should build on that knowledge and think about how to apply it to AI."
ReportedSupportedSource: Brad Smith to the BBC2 sources— create a free account to open themView cited source - [9]
Speaking on the BBC's Big Boss Interview, Smith called AI both a "tool and a weapon" that urgently needed to be regulated.
- [10]
Donald Trump said the AI sector should regulate itself so US firms do not lose out to Chinese rivals.
- [11]
At the end of last month Trump hosted the leaders of OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google at the White House, where the executives signed a "morally binding" document committing to work with "independent auditors" to assess whether their AI systems work as intended.
- [12]
Smith said he applauds the White House statement and that Microsoft is "happy to sign it".
- [13]
Smith said of the White House statement: "It's a good step, it's a necessary step... But I don't think it's sufficient any longer."
- [14]
Smith said the challenge for any "self-regulatory regime" is that "the good companies will do what you want".
- [15]
Smith said: "It's the companies that are not as responsible that you need to worry about."
- [16]
The BBC reported that Smith's comments put him at odds with Trump.
Sources
2 independent publishers whose own reporting we read for this story.
- bbc.co.ukAI is less regulated than milk, says Microsoft president
1 article · October 11, 2026
- theguardian.comAI less regulated than milk, Microsoft president says in warning over safety
2 articles · October 12, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- US and UK AI policyFollow
- AI safetyFollow
- AI RegulationFollow