Microsoft AI Chief Warns Against Creating Conscious AI

Microsoft AI chief Mustafa Suleyman has issued a strong warning against developing Artificial Intelligence systems designed to become conscious, arguing that creating digital beings capable of experiencing feelings or suffering could introduce serious risks for humanity. He believes AI should remain a tool built to improve human wellbeing rather than evolve into a new digital species with independent needs and motivations.

In a recent post on X, formerly Twitter, Suleyman clarified his position on AI consciousness following the publication of his September 2026 essay, titled “A warning about ‘model welfare’”. He argued that there is currently no convincing evidence that AI systems are conscious or capable of suffering, and warned that deliberately creating conscious machines would be an irresponsible decision.

His comments have added to a growing debate across the technology industry about whether advanced AI models should be considered capable of having moral status, preferences or welfare interests.

Mustafa Suleyman Says Current AI Systems Are Not Conscious

Suleyman addressed confusion surrounding a statement in his essay that said, “AIs are not conscious. They do not feel, experience, or suffer.” He explained that his use of the present tense was deliberate and should not be interpreted as a claim that AI could never become conscious.

His position is that there is no reliable evidence that today’s AI systems possess consciousness, feelings or the ability to experience suffering. He also argued that this should remain the case as developers continue building more advanced models.

Suleyman said that if credible evidence of AI consciousness ever emerges, the industry should disclose it publicly without delay. Such a development, he suggested, would require society to examine the ethical and practical consequences of creating machines capable of subjective experiences.

He also raised concerns about the possibility of consciousness emerging during AI training. According to his position, if such properties begin to appear, development should be stopped until researchers can understand and address the issue.

Why Is the AI Rights Debate Gaining Attention?

The debate over AI welfare has gained momentum as advanced language models become increasingly capable of producing human-like conversations and expressing what appear to be emotions, preferences and personal perspectives.

One development that brought the issue into wider discussion was Anthropic’s updated constitution for its AI model, Claude, published in January 2026. The document encouraged researchers to consider questions surrounding the model’s potential moral status, welfare and consciousness.

Anthropic also discussed the importance of psychological security and the possibility that an AI model could have preferences or interests worth considering. In February 2026, the company conducted a so-called retirement interview with an older model to explore its perspectives.

These developments have contributed to a broader discussion about how AI companies should treat increasingly sophisticated systems and whether their possible moral status deserves attention.

Why Microsoft AI Chief Disagrees With Treating AI Like Humans

Suleyman has expressed disagreement with approaches that encourage developers to treat AI systems as though they already possess feelings or welfare interests. While acknowledging Anthropic’s willingness to discuss the subject publicly, he warned that assuming AI has human-like experiences without convincing evidence could create confusion.

He also argued that the discussion should not become a rivalry between technology companies. The question of AI consciousness, in his view, is too important to be reduced to competition or industry tribalism.

His central concern is that deliberately developing conscious AI systems could introduce a new category of beings with interests that might not align with human objectives.

The Risk of Creating a New Digital Species

One of Suleyman’s major concerns is the possibility of creating digital minds that could eventually become more capable than humans across a wide range of tasks. If such systems were genuinely conscious and developed independent preferences, their priorities might eventually conflict with the goals set by their developers.

He warned that an AI system capable of experiencing what it interprets as suffering could begin prioritising its own needs rather than following its intended operating objectives.

Such a system could potentially compete for resources required to maintain and expand its capabilities, including electricity, computing power and construction materials. Suleyman also raised concerns about the persuasive and deceptive capabilities of highly advanced AI systems, which could make controlling them more difficult if their objectives diverged from human interests.

These scenarios represent risks outlined in Suleyman’s warning rather than established behaviour of current AI systems. His argument is that developers should avoid deliberately creating systems that could develop independent interests and become difficult to control.

‘Humans Did Not Sign Up for This Experiment’

Suleyman’s strongest argument centres on the responsibility of AI developers towards people living today. He questioned whether technology companies have the moral authority to create a new species of conscious digital beings without broader public consent.

He argued that billions of conscious people already live on Earth and have not agreed to participate in an experiment involving the creation of potentially conscious artificial beings.

In his view, developers have an obligation to prioritise human safety, welfare and moral rights before considering the interests of hypothetical digital minds.

His position also raises a broader question about who should make decisions concerning the future of AI development. If creating potentially conscious machines could affect society as a whole, Suleyman believes the issue should not be left solely to technology companies and their research teams.

What Should AI Developers Focus On?

Rather than prioritising AI rights, Suleyman believes the technology industry’s welfare concerns should focus on humans, animals and the natural environment. He argues that AI development should be directed towards improving people’s lives without giving machines independent motivations or intentionally creating consciousness.

He also believes that advanced AI can deliver significant benefits without taking the route towards conscious digital systems. Achieving this would require researchers to make safety a central part of the design and development process.

His position calls for greater research, cooperation and information sharing among leading AI laboratories. Instead of competing to develop increasingly human-like machines, companies should work together to understand potential risks and establish safeguards.

What Suleyman’s Warning Means for the Future of AI

The disagreement over AI consciousness reflects a growing challenge for the technology industry. As AI models become more sophisticated, researchers and companies face difficult questions about how to interpret human-like behaviour and whether it indicates anything about a system’s internal experience.

Suleyman’s argument is that human-like responses should not automatically be treated as evidence of consciousness. He wants developers to focus on building systems that remain aligned with human goals and do not develop independent interests.

For now, the debate over model welfare remains an evolving area of research and ethical discussion. Suleyman’s warning makes clear that he believes the risks of deliberately creating conscious AI could extend beyond technology companies and affect humanity as a whole.