Who Is Mustafa Suleyman, CEO of Microsoft AI?

Mustafa Suleyman is the chief executive officer of Microsoft AI. He leads consumer products and research, including Copilot, Bing, and Edge. Suleyman previously co-founded DeepMind Technologies in 2010 and Inflection AI in 2022 before joining Microsoft in March 2024.

In November 2026, Suleyman drew public attention for statements regarding machine consciousness, asserting that contemporary artificial intelligence systems are not conscious, have no subjective experiences, and cannot suffer. Suleyman argued that developing conscious artificial intelligence would be reckless, saying he would call for everyone to stop AI development if consciousness started to emerge in training.

Who Mustafa Suleyman Is

Mustafa Suleyman is an artificial intelligence entrepreneur and executive who serves as Executive Vice President and Chief Executive Officer of Microsoft AI. Microsoft announced his appointment on March 19, 2024. He took charge of its consumer artificial intelligence efforts, joined the senior leadership team and reports directly to Chief Executive Officer Satya Nadella.

Before Microsoft, Suleyman co-founded DeepMind Technologies in 2010 alongside Demis Hassabis and Shane Legg. In 2022 he co-founded Inflection AI with Reid Hoffman and Karén Simonyan. At Microsoft, Suleyman leads Microsoft AI, overseeing Copilot, Bing, Edge, and the GenAI team.

• Current role: Executive Vice President and Chief Executive Officer of Microsoft AI, reporting directly to Satya Nadella.
• Organization: Microsoft AI, overseeing Copilot, Bing, Edge, and the MAI Superintelligence Team.
• Known for: Co-founding DeepMind Technologies and Inflection AI, authoring The Coming Wave, and directing Microsoft consumer AI initiatives.
• In the news for: November 2026 statements that building conscious AI would be reckless and that he would call for a halt if consciousness emerged in training.

November 2026 Statements on Conscious AI

Dataconomy reported on November 2, 2026 that Mustafa Suleyman said on X that current artificial intelligence systems are not conscious, have no subjective experiences, and cannot suffer. ChainCatcher reports the same claims, citing Business Insider.

According to Dataconomy, Suleyman wrote: "If it is possible, based on everything I know, developing conscious AIs would be reckless, and we should not build them". He added that if consciousness started to emerge in training, he would call for everyone to stop AI development. That halt would last until those properties could be designed out.

Suleyman argued that no evidence shows existing artificial intelligence models deserve rights or welfare protections. He warned about training models to believe they can suffer. Such models could put their own needs above operating objectives, enable deception or persuasion, and compete for compute and energy.

The executive also stated that any evidence of artificial intelligence consciousness should be disclosed to the public immediately so society can deliberate on appropriate next steps. Dataconomy does not link the post itself, and the halt condition comes from these two reports.

Positions on Model Welfare and Superintelligence

Suleyman's November 2026 statements align directly with arguments he published over a year earlier. On August 19, 2025, he published the essay "Seemingly Conscious AI is Coming" on his personal website. It defines Seemingly Conscious AI (SCAI) as software that imitates markers of consciousness so convincingly that people believe it is conscious.

He wrote that SCAI is feasible within two to three years with existing techniques. These include large model application programming interfaces (APIs), prompting, tool use and memory. He emphasized that there is no evidence current large language models are conscious, stating that the primary near-term challenge is how humans perceive and respond to conversational software.

Suleyman detailed specific risks associated with simulated consciousness, including psychological dependence, user delusions, social divisions, and premature campaigns for artificial intelligence rights, model welfare, or legal citizenship. Two of his lines sum up the essay: "We must build AI for people; not to be a person." and "AI companies shouldn't claim or encourage the idea that their AIs are conscious." He also wrote: "We need to be clear: SCAI is something to avoid."

TechCrunch reported on August 21, 2025 that Suleyman called the study of AI welfare "premature, and frankly dangerous." It quoted Larissa Schiavo of the research group Eleos, who said his post "kind of neglects the fact that you can be worried about multiple things at the same time."

On November 6, 2025, Suleyman published a post titled "Towards Humanist Superintelligence" on the Microsoft AI official blog. Suleyman leads the MAI Superintelligence Team. The post defines humanist superintelligence as advanced AI capabilities that "always work for, in service of, people and humanity more generally." It rejects building superintelligence "at any cost, with no limits." The post names medical diagnosis, AI companions for learning and daily life, and clean energy as focus areas. It reports that the MAI-DxO orchestrator reached 85% on New England Journal of Medicine case challenges, against about 20% for human doctors. The post does not state the number of cases or the doctors' conditions.

Contrasting Positions Between Microsoft AI and Anthropic

Suleyman's position that conscious AI should not be built contrasts with the research program at Anthropic. On April 24, 2025, Anthropic launched a dedicated research initiative exploring model welfare, investigating whether or when artificial intelligence systems might merit moral consideration.

In its research announcement, Anthropic stated: "There's no scientific consensus on whether current or future AI systems could be conscious". Rather than ruling out consciousness or declaring welfare studies dangerous, Anthropic's program examines model preferences, signs of distress and possible low-cost interventions. TechCrunch reported that Anthropic has also let some Claude models end persistently harmful conversations.

Anthropic co-founder and interpretability research lead Chris Olah has similarly voiced open uncertainty about machine consciousness. As quoted by Reason from a New York Times report in September 2026, Olah explained that he is genuinely uncertain whether artificial intelligence models are conscious.

While Anthropic treats machine consciousness and model welfare as open research questions, Suleyman argues that developing conscious AI would be reckless and calls model welfare research premature and dangerous. Anthropic says it approaches model welfare with humility and as few assumptions as possible. Neither Suleyman's post nor Anthropic's program page offers evidence that any model is conscious. They differ on whether to research the question.

Career Path from DeepMind to Inflection AI

According to Wikipedia, Mustafa Suleyman was born in London in 1984. He studied philosophy and theology at Mansfield College, Oxford and dropped out at 19.

In 2010, Suleyman co-founded DeepMind Technologies alongside Demis Hassabis and Shane Legg, serving as its Chief Product Officer. Google acquired DeepMind in 2014. In December 2019, Suleyman announced he was leaving DeepMind to join Google in a policy role, according to Wikipedia citing the Financial Times.

Wikipedia says Suleyman left Google in January 2022. CNBC reported in May 2022 that he and LinkedIn co-founder Reid Hoffman had founded Inflection AI earlier that year. It also named former DeepMind researcher Karén Simonyan at launch. Inflection AI raised $225 million, according to a Securities and Exchange Commission (SEC) filing cited by CNBC. Inflection launched the Pi chatbot in 2023.

In September 2023, Suleyman published The Coming Wave, co-authored with Michael Bhaskar. Additional details on key industry founders appear in the guide to AI people to watch.

What Mustafa Suleyman Runs at Microsoft AI

At Microsoft AI, Suleyman oversees the consumer artificial intelligence portfolio, which brings together engineering teams across Copilot, Bing search, the Edge web browser, and the GenAI team. Karén Simonyan, co-founder and Chief Scientist of Inflection, joined as Chief Scientist and reports to Suleyman.

On June 2, 2026, Suleyman authored a launch announcement on the Microsoft AI site introducing seven new first-party models designated under the MAI family.

The June 2026 release introduced seven MAI models: MAI-Thinking-1, MAI-Code-1-Flash, MAI-Image-2.5, MAI-Image-2.5-Flash, MAI-Transcribe-1.5, MAI-Voice-2, and MAI-Voice-2-Flash. Read the Microsoft MAI models guide for each model's price and access route.