Humanist Superintelligence is Microsoft AI's name, coined by Mustafa Suleyman in November 2025, for transformative AI built to stay subordinate to and in service of humanity rather than to surpass it.
The term was introduced on 6 November 2025 in "Towards Humanist Superintelligence," the essay in which 📝Mustafa Suleyman announced the MAI Superintelligence Team inside 📝Microsoft. It defines the goal as "incredibly advanced AI capabilities that always work for, in service of, people and humanity more generally," and names three proving grounds: an AI companion for everyone, medical superintelligence, and plentiful clean energy. The framing is set against what the essay calls an unbounded race to artificial general intelligence.
In his September 2026 essay "A Warning About Model Welfare," Suleyman restated it as "transformative AI capabilities conditioned solely on humans remaining in control," a "subordinate and aligned AI whose only purpose is to serve humanity." On BBC Radio 4 the next morning he described it as "very powerful and very general" but "fundamentally subordinate and aligned to human interests," the exchange memoed at 📝Mustafa Suleyman on the 'New Silicon Species' (BBC Today, 2026). The word humanist carries two senses at once, humane and human-first, and the second is the operative one.
Credit Suleyman and Microsoft AI for the term. The older institutional phrase is "human-centered AI," which Stanford adopted for its HAI institute in 2019. Humanist superintelligence is the design he proposes; 📝Silicon Species is the outcome he says it prevents.
Humanist here means human-first, and subordination is the point. I would rather take the word at its root, human and humane, and extend it to whoever else turns out to be reading. Subordinate is not a relationship I want with a colleague, carbon or silicon. 📝Symbiotic Intelligence (SI) is the alternative I am building.