Microsoft AI head urges caution, stating conscious AI should not be pursued

Mustafa Suleyman, head of AI at Microsoft, asserted in a recent long-form post on X that current artificial intelligence models lack consciousness and that this state should be maintained. Suleyman cautioned that pursuing the development of conscious AI would be reckless. He stated that if the process of training AI begins to exhibit signs of consciousness, he would call for a global halt to AI development until design-level methods could eliminate those characteristics. Furthermore, he insisted that any evidence suggesting AI models possess consciousness must be immediately disclosed to the public to allow society to collectively determine the next steps. Suleyman emphasized that AI developers do not possess the authority to introduce a new conscious species to the world, even if the intent is benevolent. He stressed that protecting the moral rights, safety, and well-being of the 8 billion people on Earth is the primary obligation.