Satya Nadella: Superintelligence Isn’t Worth Pursuing Without Human Control & Benefits for Humanity

Although Microsoft is open about AI adoption and has been integrating its in-house AI models across its products and services, the company’s CEO Satya Nadella has openly talked out the next phase of AI, and he wants one principle kept at the center. Any pursuit of superintelligence, he says, must ultimately help humanity while keeping people in control.

Nadella wants AI benefits spread across the ecosystem

In a post on X, Nadella argued that AI’s benefits should reach countries, communities, and companies broadly. He also called for an AI ecosystem where both closed and open-source models can develop.

That approach could challenge the idea that increasingly capable AI should remain concentrated among a small group of companies. Nadella said broad representation is needed across countries, industries, academia, and the wider AI ecosystem.

He also highlighted a major concern for businesses building AI systems. Companies should retain control over their own unique knowledge instead of becoming dependent on a single model provider. According to Nadella, organizations should be able to build their own continuous learning loops. They should also have ways to embed their proprietary knowledge into models and weights they control.

Microsoft is preparing its own AI code of conduct

Nadella also welcomed more research around AI alignment and what he described as deliberate pacing. He specifically pointed to ideas such as “embedded evaluators” as part of efforts to turn AI safety principles into practical mechanisms.

The comments come as Microsoft continues developing its own MAI models. Nadella said the company plans to publish a Code of Conduct for its first-party MAI models for public consultation.

That could give researchers, businesses, and other AI developers a closer look at the principles Microsoft wants guiding its models. It also fits with Nadella’s broader argument that AI development should offer choice at every layer of the stack.

At a time when the race toward increasingly powerful AI is accelerating, Nadella’s comments put control and access alongside raw capability. In the past few days, there have been some chilling warnings coming out from the AI industry, along with some latest discoveries, suggesting that if AI is not controlled, it could have catastrophic results for humanity.