Microsoft AI CEO warns companies working on AI to stop confusing control with co-operation, Mustafa Suleyman says, ‘You can’t steer something…’

Microsoft AI CEO on containment and alignment
Mustafa Süleyman emphasized that this distinction is crucial because the AI industry often treats constraint and alignment as if they were interchangeable, even though they create very different technical and philosophical challenges. He explained that containment is about enforcing strict boundaries and restricting agency by keeping AI systems within clearly defined boundaries. Alignment, by contrast, focuses on whether these systems will act in the best interests of humanity. He argued that maintaining alignment without first establishing a strong constraint was effectively putting the cart before the horse.
ALSO READ: Prashant Tamang’s latest Instagram post was all about ‘exciting, strong crowd and unrivaled vibes’
These warnings come as Suleiman positions Microsoft as a counterweight to what he sees as reckless AI development elsewhere in the industry. In his recent article, “Towards Humanistic Superintelligence,” published on the Microsoft AI blog, he outlined a vision that prioritizes human control and domain-specific AI over unconstrained, autonomous systems. He told Bloomberg in a December interview that both containment and compliance should be treated as “red lines” that no company should cross, acknowledging that stance remains “a new position in the industry.”
At the heart of the alternative Suleiman proposes (what he calls Humanistic Superintelligence) are practical applications such as medical diagnostics and clean energy, rather than the search for general-purpose artificial general intelligence. Microsoft AI recently developed a system that achieved 85% accuracy on highly challenging case studies from the New England Journal of Medicine; for human doctors, the rate was roughly 20%. The former DeepMind co-founder, who joined Microsoft 18 months ago, argues that this domain-focused strategy can deliver superintelligence-level performance while eliminating the most serious control risks.
With a revised OpenAI deal now allowing Microsoft to pursue independent AI development, Suleiman says he has established what he calls the world’s leading superintelligence research team; It is stated that this team is clearly designed to keep people under tight control.
ALSO READ: Indian Idol Season 3 winner and Paatal Lok actor Prashant Tamang dies at 43
Mustafa Süleyman called Musk a ‘bulldozer’
Last month, Mustafa Suleyman said he was in regular contact with colleagues in the artificial intelligence field, including OpenAI’s Sam Altman, Anthropic’s Dario Amodei and Google DeepMind’s Demis Hassabis. He was asked about Elon Musk, who co-founded OpenAI but has since fallen out with Altman and is pursuing artificial intelligence through his own startup xAI. Suleiman called him “bulldozer”. Although Musk has “different kinds of values,” Solomon said: “I like it when he speaks his mind. It’s very unfiltered.” “He may be one of the greatest entrepreneurs of our generation,” Suleiman added. “He’s certainly accomplished a lot. He’s building data centers faster than anyone in the industry, and if he can achieve that, it’ll be pretty dramatic.”



