Geoffrey Hinton, Stuart Russell and other AI leaders are warning us that there are existential risks, if only due to the inevitable creation of autonomous murder machines by the military (Hinton and Russell are both worried about things beyond that, by the way). I’ll take their word for it and not only on the basis of their authority - they have very well put arguments.

OpenAI is also playing it safe to some extent, maybe to please the general public and serve the org’s own interests, or maybe they are genuinely concerned about real safety threats, who knows. At any rate their actions don’t align with your discourse.