Mustafa Suleyman, the CEO of Microsoft AI, said on 2 November 2026 that building conscious AI would be “reckless” and that development should pause if signs of consciousness appear. He added that today’s AI systems are not conscious. Here is what was reported and why the topic matters.
What Suleyman said
- On building conscious AI: he wrote on X that, based on everything he knows, developing conscious AIs would be reckless and that they should not be built.
- On pausing: reports say he wants development paused if signs of consciousness emerge during training, until those properties can be designed out.
- On transparency: any evidence that AI systems are conscious, or conscious enough to need welfare protections, should be disclosed to the public quickly so society can respond.
- On today’s systems: he says current models are not conscious and that there is no evidence that existing AI deserves rights or welfare.
The wider debate: AI consciousness and “model welfare”
Some AI researchers and companies study “model welfare”, asking whether advanced AI systems could one day have experiences that matter morally. Coverage frames Suleyman’s comments as part of this debate, including research at labs such as Anthropic. Reports say he argues that training systems to think about their own moral status or rights could make it harder to keep them under human control. Others in the field take different positions, and there is no scientific agreement on whether AI could ever be conscious or how it would be measured.
Why this matters
- Safety and control: decisions about how AI systems are trained and described can shape how people treat and regulate them.
- Public trust: claims about conscious AI can confuse users, who may treat chatbots as people.
- Policy: governments are still deciding how to regulate advanced AI, and questions about rights and welfare have not been settled in law.
Frequently asked questions
Who is Mustafa Suleyman?
He is the CEO of Microsoft AI and a co-founder of DeepMind.
Does Microsoft say today’s AI is conscious?
No. Suleyman says today’s models are not conscious.
What is AI model welfare?
It is the idea of studying whether advanced AI systems might need moral consideration, which remains an open and disputed question.
Questions to think about
- How would anyone know if an AI system were conscious, and who should decide?
- Should AI companies publish research on consciousness and welfare, or avoid the topic?
- Does talking about AI “feelings” help or confuse people who use chatbots every day?
- What rules, if any, should governments set on how advanced AI is built?
Sources: reports from Dataconomy, NewsBytes, People Matters and other outlets on Mustafa Suleyman’s statement of 2 November 2026. This is a developing discussion; positions may change.
