While tech rivals race to outsmart humans, Microsoft says its AI will stay on humanity’s side.
Microsoft has launched a new Superintelligence Team within its AI division, marking a significant step in its pursuit of what it calls “humanist superintelligence” — a vision for powerful AI that remains firmly under human control.
Announced on Thursday by Mustafa Suleyman, CEO of Microsoft AI, the new MAI Superintelligence Team will focus on developing advanced AI that works for humanity rather than competing with it.
“We are doing this to solve real concrete problems and do it in such a way that it remains grounded and controllable,” Suleyman wrote in a blog post. “We are not building an ill-defined and ethereal superintelligence; we are building a practical technology explicitly designed only to serve humanity.”
The focus is on creating systems that are “problem-oriented and tend towards the domain specific,” rather than on a general-purpose AI that can surpass human performance on all tasks — a concept known as AGI, which is a key focus for Microsoft’s longtime partner, OpenAI.
Rejecting the AGI race narrative
Microsoft’s announcement comes just months after Meta launched its own Superintelligence Lab, led by Scale AI founder Alexandr Wang, and follows similar moves by other players, such as Safe Superintelligence Inc., founded by former OpenAI chief scientist Ilya Sutskever.
However, Suleyman indicated the team is looking beyond the competitive frenzy.
In his post, he explained that Microsoft “reject[s] narratives about a race to AGI, and instead see[s] it as part of a wider and deeply human endeavour to improve our lives and future prospects.” Suleyman acknowledged the profound responsibility involved, writing, “We feel a deep responsibility to get this right.”
The MAI Superintelligence Team will be led by Suleyman alongside Microsoft AI Chief Scientist Karén Simonyan and other core leaders, aiming to be the “world’s best place to research and build AI, bar none.”
Focus areas: Health, AI companions, and clean energy
Suleyman’s vision for “humanist superintelligence” centers on solving real-world challenges, rather than creating machines with unchecked autonomy. He cited three areas where the technology could make an early impact: healthcare, AI companions, and clean energy.
In healthcare, Microsoft aims to develop medical superintelligence capable of expert-level diagnostics.
“We’ll have expert level performance at the full range of diagnostics, alongside highly capable planning and prediction in operational clinical settings,” Suleyman wrote.
He cited Microsoft’s recent work on a medical orchestrator, MAI-DxO, which achieved an 85% success rate on a challenging New England Journal of Medicine Case Challenge, far exceeding the approximately 20% max rate for human doctors.
For AI companions, the company envisions personalized digital assistants that can help users learn, stay productive, and manage everyday life while keeping human relationships at the center. And in energy, Microsoft believes AI will help accelerate breakthroughs in battery storage, materials science, and renewable energy systems.
The overall mission is to ensure the most advanced forms of AI “keep humanity in control” while tackling the world’s most pressing challenges.
Suleyman emphasized the criticality of this approach, saying, “Creating superintelligence is one thing; but creating provable, robust containment and alignment alongside it is the urgent challenge facing humanity in the 21st century.”
Earlier this week, Suleyman dismissed the idea of conscious AI, saying that consciousness is reserved for humans.