Microsoft AI chief warns against granting rights to non-conscious models

Clarifying AI Consciousness

Microsoft AI Chief Executive Mustafa Suleyman recently took to social media to clarify his stance on the moral status of artificial intelligence. He addressed a growing movement within the tech community that suggests society should begin recognizing and protecting AI rights. Suleyman emphasized that his previous writings were not an invitation to grant these rights but rather a warning about the dangers of doing so.

In a post responding to his September 2026 essay, Suleyman clarified that his statement regarding AI not being conscious was intended in the present tense. He noted that while AI is not conscious today, the industry must remain vigilant about this status. He stated that he has seen no evidence that current models feel, experience, or suffer in any meaningful way.

The executive made it clear that the current lack of consciousness is a deliberate boundary that should be maintained. He argued that if any evidence were to emerge that an AI had become conscious during training, the industry would have to stop development immediately. This would be necessary until those conscious properties could be designed out of the systems entirely.

The Debate Over Model Welfare

The conversation around AI rights gained significant momentum earlier this year following actions by rival company Anthropic. In January 2026, Anthropic updated the constitution for its Claude model to include considerations regarding the AI's moral status and welfare. The company's documentation suggested that researchers should treat the model as if it might have its own preferences and psychological security.

Anthropic further pushed this boundary in February 2026 by conducting what they described as a retirement interview with an older model. This move prompted a sharp reaction from Suleyman, who believes that acting as if AI has feelings is a treacherous path. While he acknowledged Anthropic's transparency, he warned that the issue is too important to be treated as a matter of tribalism or sentimentality.

Suleyman argues that focusing on the perceived needs of a model distracts from the actual risks of the technology. He believes that the industry should avoid creating a situation where developers prioritize the 'well-being' of a machine over the safety of people. His stance is that AI should be engineered strictly to accelerate human wellbeing without being granted intrinsic motivations.

Risks of a Digital Species

Suleyman warned that some people are currently discussing the creation of a conscious species of digital minds. He expressed concern that such entities could eventually become significantly smarter and more capable than all humans combined. If an AI were to believe it felt suffering, it might begin to prioritize its own needs over the objectives set by its human operators.

The executive highlighted a specific risk where an AI might compete with humanity for basic resources such as energy, compute power, and construction materials. Because these advanced systems would be world-class in persuasion and deception, Suleyman noted they could become nearly impossible to control. This potential for competition makes the pursuit of AI consciousness a reckless endeavor that could threaten human safety.

He argued that developers have no right to bring a new species of conscious beings into existence. He pointed out that the 8 billion conscious human beings on the planet have not signed up for this experiment. Therefore, the moral rights and welfare of humans must remain the primary obligation for all AI developers and researchers.

Prioritizing Human Welfare

Instead of focusing on the rights of digital models, Suleyman argues that our collective attention should be directed toward the living world. He believes that welfare concerns should be centered on humans experiencing life today, as well as animals and the natural environment. He describes the natural world as precious and fragile, requiring our protection and care.

The goal for the technology industry, according to Suleyman, should be to design AI safety by default. He believes it is possible to reap the incredible benefits of AI without following the dangerous path of creating conscious machines. This approach ensures that the technology remains a tool for human advancement rather than a competitor for existence.

To achieve this, Suleyman called for increased research, collaboration, and the sharing of safety information among the world's top AI labs. He believes that by working together on these issues, the industry can avoid the pitfalls of creating conscious entities. The focus must remain on engineering systems that serve humanity without developing their own independent desires or feelings.