On December 12, 2025, a report from Epoch Times revealed that Mustafa Suleyman, the Head of Artificial Intelligence (AI) at the American tech giant Microsoft, stated that if super-intelligence poses a threat to humanity, they will immediately halt relevant development work.
In an interview with Bloomberg on Thursday, December 11, Mustafa Suleyman, the Consumer AI Chief at Microsoft, explicitly stated their dedication to creating a super-intelligent system that aligns with human interests. He emphasized, “We will not develop a system that could potentially get out of control, and we will never let it continue.”
He added that this should be an industry consensus, a universal standpoint.
Suleyman’s commitment highlights the ethical considerations that tech giants face when pursuing breakthroughs in AI technology. Competitors like OpenAI and Anthropic have previously made statements similar to Microsoft.
In early 2024, Microsoft acquired the intellectual property of Inflection AI, a startup founded by Suleyman, along with most of its employees, after which he announced his joining of Microsoft. Microsoft has tasked Suleyman with building products that can rival the best AI tools in the industry. In November 2025, Suleyman announced the establishment of the MAI Superintelligence Team, clearly stating their intention to develop practical technologies designed and serviced for humanity.
When announcing this superintelligence project, Suleyman mentioned that success can only be achieved by ensuring that AI systems are designed to serve humanity.
His team will focus on solving real-world problems in fields such as medical diagnosis and education, rather than focusing on “unclear and abstract superintelligence.”
The team’s primary goal is to develop AI tools with capabilities far beyond human capacity in specific areas like medical diagnosis. Suleyman wrote in a blog post, “We will achieve expert-level performance in comprehensive diagnosis, while possessing high levels of planning and predictive abilities in clinical operational environments.”
Currently, discussions around superintelligence are still primarily at the academic level. While AI systems like ChatGPT are able to interact with humans, there is still a gap in meeting consumer and business expectations.
Suleyman admits that the current AI technology is still evolving. He mentioned that the AI feature in Microsoft’s Copilot consumer assistant is “not always accurate” and is currently in the development stage.