A senior executive at a tech giant who personally helped push frontier models to market has publicly drawn a red line around "AI consciousness." On November 2, it was reported that Microsoft AI chief Mustafa Suleyman published a long post on X, stating that current models are not conscious and that humanity should keep it that way forever. As a co-founder of DeepMind and the head of Microsoft's AI division, the weight of this statement far exceeds the private concerns of an ordinary engineer.

From "Not Conscious" to "Should Not Be Conscious"

Suleyman's wording goes far beyond a "description of the status quo." He wrote: as long as building conscious AI is technically feasible, then doing so is itself reckless, and we should not develop such things. Even more decisively, if signs of consciousness emerge during training, he would call for the entire industry to pause immediately until those traits can be completely designed out. In other words, for him consciousness is not a scientific question of "whether it can be achieved," but an ethical red line that "must be prohibited." At a time when regulators in various countries are still chasing model capabilities with legislation, he is effectively drawing the line before capability.

Who Agreed to This Experiment?

On the "disclosure" front, his position is equally clear: if any evidence appears that an AI is conscious, it should immediately be made public so that society as a whole can decide the next step. His reasoning rests on the boundaries of responsibility—developers simply have no right to bring a new conscious species into the world, even if their original intent puts well-being first. "The 8 billion people on Earth did not agree to participate in this experiment." Protecting people's moral rights, safety, and well-being is the first priority. While the industry is still competing over parameters and levels of intelligence, these remarks drag the conversation back to the simplest and sharpest question: what exactly should we build?