Mustafa Suleyman, CEO of Microsoft AI and former DeepMind executive, warned on BBC Radio 4 that rapidly improving AI could become "uncontrollable" without stronger regulation and industry guardrails. He described public concern as "healthy and necessary" and forecast "outrageously exponential" advances over the next five years. Microsoft research of over 200,000 Copilot interactions highlighted roles already affected, and Suleyman warned that many predictable white‑collar tasks could soon be automated.
Microsoft AI Chief Warns AI Could Become 'Uncontrollable' — Urges Urgent Guardrails
CEO of Microsoft AI, Mustafa Suleyman, said that fear over the future of AI is “healthy and necessary” (Doug Peters/PA) (PA Wire)
Mustafa Suleyman, CEO of Microsoft AI and co‑founder and former head of applied AI at DeepMind, warned on BBC Radio 4’s Today programme that rapidly advancing artificial intelligence could become "uncontrollable" within years if regulators and industry do not put stronger guardrails in place.
Why He’s Concerned
Speaking as a guest editor on the programme, Suleyman described public worry about AI as "healthy and necessary," adding: "I honestly think that if you’re not a little bit afraid at this moment then you’re not paying attention." He predicted that progress in the next five years will be "outrageously exponential," raising the prospect of systems that could outperform humans across nearly all tasks.
"A system like that would almost certainly not be controllable. We have to declare our belief in a humanist super intelligence, one that is always aligned to human interests."
Potential Social and Economic Impact
Suleyman warned that if advanced systems cannot be controlled, "it isn’t going to be on our side. It’s going to overwhelm us." He stressed that many present AI advances are "fundamentally labour‑replacing technologies," capable of performing tasks previously done by white‑collar workers.
Research from Microsoft AI earlier this year revealed which jobs are most at risk from AI automation (Niall Carson/PA) (PA Archive)
Microsoft AI research analysing more than 200,000 interactions with its Copilot generative AI found that roles such as interpreters and translators, historians, mathematicians and proofreaders were frequently performed by the tool. Suleyman pointed to call‑centre agents as an example of workers already being affected and predicted that paralegals, junior accountants and general‑purpose project managers are likely to follow.
What He’s Calling For
Suleyman urged governments, regulators and companies to implement stronger guardrails, public oversight and safety standards to ensure AI development remains aligned with human interests. He advocated for a "humanist super intelligence" — advanced systems intentionally designed to serve and protect human values and priorities.
While highlighting the efficiency gains AI can deliver, he also emphasised the need for public debate and policy action to manage risks, protect jobs, and ensure benefits are widely shared.
Context: Suleyman’s comments come amid growing international discussion about AI regulation, workforce displacement, and the balance between innovation and safety.
