Artificial Intelligence

Microsoft AI Boss Warns of Growing Drive to Build Human-Replacing Superintelligence

Published NOV 12 2025

Story

Above: Microsoft AI CEO Mustafa Suleyman at an event commemorating Microsoft’s 50th anniversary in Redmond, Washington, U.S., on April 4, 2025. Image credit: David Ryder/Bloomberg/Getty Images

The Facts

Mustafa Suleyman, the CEO of Microsoft AI, has claimed to have seen “many indications” that other actors in the tech sector are actively pursuing AI systems capable of “replacing or threatening our species,” during an interview with Semafor.
ControlAI on X
Semafor on Youtube
Suleyman described the danger as an AI system with beyond-human levels of intelligence capable of self-improvement, complete autonomy and independent goal-setting, stating he can only “hope” AI alignment prevents such systems from turning against humanity.
Semafor on Youtube
ControlAI on X
According to Suleyman, Microsoft, in contrast, will only pursue a “humanist super intelligence” model centered on “control, containment and alignment,” as well as human wellbeing, to avoid a “transhumanist moment” that would “leave humans behind.”
Semafor on Youtube
His remarks follow Microsoft’s launch of an AI superintelligence team in early November, which Suleyman said will focus on solving “real concrete problems” in a “grounded and controllable” way, which includes creating a medical superintelligence, AI companions and achieving “plentiful clean energy.”
Microsoft AI
The team will focus on in-house research and development to make Microsoft “self-sufficient in AI,” marking a departure from the company’s previous reliance on OpenAI, following an agreement between the two that allows Microsoft to pursue AGI independently.
Windows Central
Business Insider
The Official Microsoft Blog
Suleyman joined Microsoft in 2024. Before joining the company, Suleyman founded the startup Inflection AI and co-founded DeepMind, an AI lab that Google acquired in 2014, where Suleyman also worked for several years.
CNBC
The Wall Street Journal

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Techno-optimist narrative

AI superintelligence, if managed correctly, can revolutionise the human experience. From its promise of revitalizing the global economy to its potential to reshape medicine and energy, AI superintelligence can build a better tomorrow, but only if developed in a way that is safe, controllable and aligned with human interests.

Microsoft AI

Mustafa Suleyman on X

Techno-skeptic narrative

While innovative AI may bring many benefits to everyday life, the risks associated with AI superintelligence outweigh any potential rewards. From the threat of human economic obsolescence to extinction, an AI superintelligence could be catastrophic for humanity, which is why international parties must agree to prohibit its development.

Future of Life Institute