Microsoft’s AI chief says superintelligence is coming—but your job is safe

Mustafa Suleyman, the CEO of Microsoft AI, has a clear message: superintelligence is on the horizon, but it won’t replace you at work. In a wide-ranging interview, Suleyman discussed the company’s evolving relationship with OpenAI, its push to build its own frontier models, and why he thinks calling AI “alive” is a dangerous mistake.

Suleyman, who recently shifted focus from consumer products to leading a new Superintelligence team, said Microsoft is now training its own models from scratch—a move driven by the realization that long-term dependence on a third party for the most valuable technology of all time wasn’t sustainable. “Superintelligence is coming. It’s just around the corner,” he said. “There’s no way we could be structurally dependent on someone else for that forever.”

He pushed back on fears that AI will wipe out white-collar jobs, clarifying that his earlier comments about automating tasks within 12 to 18 months were misunderstood. “Tasks are not jobs,” he explained. “Sending an email, writing a PowerPoint—those sub-tasks will be automated. That doesn’t mean the role disappears. It frees people up for more creative, judgment-focused work.”

On the topic of consciousness, Suleyman was blunt. He criticized Anthropic for speculating in its training manual that Claude might have feelings or rights. “It’s dangerous to project potential rights onto tools,” he said. “These models don’t experience suffering. They’re controllable, accountable tools that should serve humanity—not the other way around.”

Microsoft also announced a new partnership with Mayo Clinic to co-train a health-focused foundation model, part of Suleyman’s long-standing belief that technology’s purpose is to make people healthier and happier. “If it doesn’t do that,” he said, “people will reject it—and they’ll be right to.”