Microsoft is setting up a new artificial intelligence division that aims to create machines capable of outperforming humans in specific areas, but not in ways that threaten them. The company says its goal is to create “humanistic superintelligence,” or AI that enhances human well-being rather than replacing it.

The initiative, called the MAI Superintelligence Team, is led by Mustafa Suleiman, Microsoft’s AI chief and DeepMind co-founder. In an interview with Reuters, Suleiman said the project would focus on developing AI that provides “superhuman” abilities in defined domains, starting with medical diagnosis.

According to him, Microsoft envisions AI that can reason through complex problems and make discoveries that humans alone cannot reach – but within ethical boundaries. “Humanism requires us to always ask the question: Does this technology serve human interests?” Suleiman described the company’s approach to AI as enhancing humanity rather than endangering it.

Humanistic superintelligence, not generic AI

Unlike other companies pursuing artificial general intelligence (AGI) – systems capable of mastering any task performed by a human – Microsoft’s new team is not pursuing the idea of ​​an “infinitely capable generalist” machine. Suleiman said he doubted that such systems could be controlled if they were ever built. Instead, the focus will be on expert AI models that can tackle major real-world challenges like battery storage or drug discovery, while posing “virtually no existential risk.”

This approach is similar in ambition to the projects Meta Platform and SAFE Superintelligence Inc., but Microsoft’s framework differentiates it by prioritizing human outcomes. The company hopes that these efforts will not only transform industries but also increase human life expectancy by solving complex medical problems with AI precision.

Medical diagnostics is one of the areas where Suleiman believes AI can achieve transformative results within a short period of time. He revealed that Microsoft now has “its sights on medical superintelligence in the next two to three years.”

The team plans to develop AI that is able to detect and prevent diseases long before symptoms appear — a move that, as Suleiman says, “could extend our life expectancy and give everyone more healthy years.”

This direction mirrors his earlier work at DeepMind, where AI models like AlphaFold revolutionized molecular biology by predicting protein structures. Microsoft hopes to build on that legacy and achieve similar advancements in healthcare and life sciences.

Recruiting top talent and investing heavily

To power this ambitious effort, Microsoft plans to invest “a lot of money” in the MAI superintelligence team. Suleiman said the company will recruit top AI scientists from around the world, as well as bring in experienced researchers from within Microsoft. Karen Simonyan will serve as the team’s chief scientist.

While competitors like Meta have reportedly offered signing bonuses of up to $100 million to attract AI researchers, Suleiman declined to comment on whether Microsoft would match such offers. However, he made it clear that Microsoft AI will continue to hire aggressively as part of its long-term strategy.