What Nadella said
Microsoft chief executive Satya Nadella said on Sunday that the company welcomes the “deliberate pacing needed to get alignment right”, and that Microsoft would publish a Code of Conduct for its first-party MAI models on Monday, 14 September, for public consultation. He made the announcement in a post on X, which was quoted by Unite.AI and A News.
“Any pursuit of superintelligence has to be grounded in the core principle that if the AI we build is not helping humanity and under human control, it’s not worth pursuing,” Nadella wrote.
He also welcomed the idea of “embedded evaluators”, independent safety reviewers working inside the labs, one of the safeguards Anthropic chief executive Dario Amodei proposed in his essay calling on frontier developers to slow down.
A code for Microsoft’s own models
Nadella set out Microsoft’s approach in three parts: broad access and choice at every layer of the AI stack, control for businesses over their own learning loops and models, and the Code of Conduct underlying Microsoft’s first-party MAI models. “For firms, it’s imperative that they retain full control over their unique and tacit knowledge,” he wrote, according to the Free Press Journal.
Unite.AI reported that it will be the first time Microsoft has published behavioural standards for its own models as a document. Microsoft has not said how long the consultation will run or how responses will be collected, and at the time of publication, early on Monday in Europe, the document had not appeared.
Not a handful of companies
The post also carried a warning. Governance “cannot be controlled by a handful of entities, but must have broad representation across the ecosystem, countries, and fields, including academia,” Nadella wrote, adding that a healthy frontier needs “both closed and open-source models”.
That lands the same weekend as a report in The Information that Anthropic, OpenAI and Google have held working-group meetings since July about an industry-led standards body for frontier AI. According to investingLive’s account of that report, Google DeepMind chief executive Demis Hassabis proposed on 14 July a US-based Frontier AI Standards Body modelled on FINRA, the self-regulatory organisation for American broker-dealers, and OpenAI chief executive Sam Altman told staff that without government backing the big labs would have to build such a body themselves. Microsoft was not named among the participants.
Where the pacing debate stands
Nadella’s endorsement follows Altman’s statement on Saturday that OpenAI would also commit to independent evaluators with employee-like access. The response in Washington has been cooler: White House AI adviser David Sacks said the labs are already free to pace the frontier for business reasons, and House Speaker Mike Johnson said Congress would not lead on AI safety legislation.
Microsoft’s position matters beyond its own models. It is one of OpenAI’s largest investors, and its MAI models make it a frontier developer in its own right, so a code that binds MAI tells the market something about how one of the biggest buyers and builders of AI intends to behave.
What to watch
The document. A code of conduct can be a list of principles or a set of testable commitments. Whether Microsoft’s names release conditions, evaluation thresholds or outside reviewers, and whether the consultation gives outsiders a real say, will show how far its endorsement of pacing goes.
