Microsoft AI CEO Suleiman: AI models do not have consciousness and should not develop conscious AI

According to a report by Phoenix Technology citing Business Insider, on November 2, Beijing time, Microsoft AI CEO Mustafa Suleyman stated on X that based on the information he has, current AI models do not possess consciousness, and developers should ensure that this situation continues.

Suleyman stated that if developing conscious AI becomes possible, it would be reckless to do so, and such AI should not be developed. If signs of consciousness begin to emerge during the training process, he would call for everyone to pause AI development until these characteristics can be designed out. He also mentioned that if there is any evidence that AI models have developed consciousness, it must be disclosed to the public immediately so that society can collectively decide on the next steps.

Suleyman believes that AI developers have no right to bring a new conscious species into this world; the primary obligation is to protect the moral rights, safety, and well-being of the approximately 8 billion conscious humans on Earth. The report also noted that discussions in the tech community about whether AI could potentially gain consciousness are heating up. Last April, Anthropic launched a research initiative to investigate and prepare for issues related to model welfare and to discuss whether developers should focus on the potential consciousness of AI models and the experiences they may possess.