Microsoft CEO Satya Nadella Backs Slower AI Development, Warns Against Concentration of Power

Microsoft CEO Satya Nadella has called for a more deliberate approach to artificial intelligence development, saying the pursuit of increasingly capable systems must remain focused on human benefit and control.

In a post on X, Nadella addressed the growing debate over whether AI development should slow down to allow researchers and companies more time to address safety and alignment concerns. His comments also offered a broader warning to industry leaders, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, about the risks of allowing advanced AI to become concentrated among a small group of companies.

Nadella said the pursuit of superintelligence is only worthwhile if the resulting systems help humanity and remain under human control.

Nadella calls for deliberate AI development

The Microsoft chief said research, careful evaluation and deliberate pacing have a role in ensuring that AI systems are developed with alignment as a central design objective.

He also backed ideas such as "embedded evaluators" and other mechanisms intended to turn AI safety principles into practical safeguards rather than relying solely on broad commitments.

For Nadella, however, slowing down does not mean stopping progress. He argued that the benefits of AI should continue to expand and reach countries, communities and companies more broadly.

He said a healthy frontier AI ecosystem should allow both closed and open-source models to develop.

Companies should control their own AI knowledge

A key part of Nadella's argument focused on businesses maintaining control over their own knowledge and AI systems.

He said companies should be able to build their own continuous learning processes without becoming dependent on a single model provider.

"Every organization should be able to build its own continuous learning loop… without becoming dependent on any one model provider," Nadella wrote.

He argued that organisations should also have the ability to incorporate their proprietary knowledge into models and model weights they control.

The position reflects a broader concern for enterprises adopting AI services from a small number of major providers. Businesses increasingly rely on external models for software development, customer service, research and other operations, making control over proprietary information an important consideration.

Nadella warns against AI power being concentrated

Nadella also cautioned that advanced AI should not be controlled by only a handful of companies.

He called for broader representation across countries, industries and academic institutions as AI capabilities continue to advance.

His comments come as Microsoft maintains a close relationship with OpenAI while also developing its own MAI family of models. Microsoft has invested heavily in AI infrastructure and provides access to multiple models through its broader AI platform.

Nadella said Microsoft's approach is to provide choice and access across different layers of the AI technology stack while giving enterprises greater control over their own learning processes and models.

Microsoft plans public consultation on MAI rules

Nadella also announced that Microsoft plans to publish a "Code of Conduct" covering its first-party MAI models for public consultation.

The proposed framework is intended to set out the principles guiding Microsoft's own AI systems.

Nadella's comments therefore combine two positions in the AI debate: continued investment in increasingly capable systems, alongside greater caution around safety, control and concentration of AI capabilities.

His central argument is that AI development should not be measured only by how quickly models become more capable. Companies also need mechanisms that allow people and organisations to maintain control over how those systems learn, operate and are used.