Microsoft Chairman and CEO Satya Nadella has said that any effort to develop superintelligence must be based on the idea that AI should serve humanity and remain under human oversight.
In a post on X, Nadella said that if AI does not contribute to human well-being or cannot remain under human control, then pursuing it is “not worth pursuing”. He also stressed the need to speed up the distribution of AI’s benefits across nations, communities and businesses.
BROAD ACCESS TO AI
Nadella said the AI ecosystem should support both proprietary and open-source models. He argued that organisations should maintain complete control over their own unique and tacit knowledge instead of becoming reliant on one AI model provider.
“Every organisation should be able to build its own continuous learning loop/hill climbing machine,” Nadella said. He added that companies should have the ability to incorporate their own knowledge into models and weights that remain under their control.
He said Microsoft supports the research, attention and “deliberate pacing” required to ensure AI alignment is properly built into the technology from the outset. Nadella also backed concepts such as “embedded evaluators” and wider efforts aimed at creating practical mechanisms for implementing AI safety principles.
NO CONTROL BY A HANDFUL OF ENTITIES
Nadella emphasised that advanced AI development should not be concentrated in the hands of a limited number of organisations. He called for wider participation across the AI ecosystem, different countries and various fields, including academia.
“The key is that this cannot be controlled by a handful of entities,” he said.
Nadella described Microsoft's strategy as providing “broad access and choice at every layer of the AI stack”, while ensuring enterprises retain control over their learning processes and models.
He also announced that Microsoft would release a “Code of Conduct” for its first-party MAI models for public consultation the following day.
