Microsoft AI chief is open to AI rights if convincing evidence emerges

Mustafa Suleyman, Microsoft’s head of AI, is willing to reconsider his position on AI rights if evidence emerges that such systems are conscious. For now, he sees no evidence that models are conscious or capable of suffering, and believes conscious AI should not be created. He clarified his position on X after publishing an essay on model welfare.

Microsoft AI chief allows for rights for conscious AI

What Suleyman considers evidence

In his essay, Suleyman stated categorically that AI is not conscious and cannot feel, experience or suffer. After it was published, he noted that many users on X consider the question of AI rights a serious one.

Suleyman explained that he is not claiming AI could never become conscious. His position is that current systems are not conscious — and, in his view, should remain that way. He has seen no data confirming AI consciousness or the need to care about AI welfare, but is willing to change his mind if such data emerges.

Why the debate continues

The question of AI consciousness remains contested: there is no agreement even on whether the current model architecture can support experiences comparable to human ones. At the same time, some industry figures, including Anthropic, allow that language models may already be conscious. The debate is fueled, according to the original article’s author, by fears shaped by science fiction, though some people in the industry take such scenarios seriously.

Suleyman opposes creating intelligent artificial beings. He considers developing conscious AI reckless and believes developers have no right to introduce a new kind of conscious being into the world, even if someone wants to put AI welfare first.

Dataist’s take: Suleyman leaves room to reconsider his views if new evidence emerges, but for now opposes creating conscious AI. The question of what evidence could confirm a model’s consciousness remains contested.