Microsoft on Thursday announced the formation of its new MAI Superintelligence Team, a major initiative aimed at developing artificial-intelligence systems that surpass human capabilities in specific fields, starting with medical diagnosis.
Led by AI chief Mustafa Suleyman, the project marks a strategic shift from broad-based general-purpose AI to specialized “humanist superintelligence” — technology designed to solve defined problems with human-centric safeguards.
Unlike some competitors chasing fully autonomous AI, Microsoft says it will focus on domain-specific models that offer super-human performance while minimizing the risk of uncontrolled systems. Suleyman gave examples including battery storage, molecular development and disease diagnosis.
The initiative will begin with medical diagnostics, which Microsoft believes offers a clear path to “medical superintelligence” within two to three years.
Microsoft plans to invest heavily in the effort and populate the new team with both existing researchers and new talent. Karen Simonyan will serve as chief scientist.
The company said it is not simply chasing an all-purpose AI, but rather “humanist superintelligence” — AI that is powerful within defined problems and aligned with human interests.