I don't think ML is substantially different from other technologies that we have deployed that we didn't completely understand: early atomic weapons, early molecular biology, cars, and more. In each case, by deploying a bit with caution, we learned important lessons that allowed us to continue. Rather than sticking our head in the ground and remaining ignorant. In each case there were people calling for the deployments to stop. Of all those, I think only atomic weapons ever represented an existential risk to humanity- and we figured that out within a decade or to and set up political structures to prevent it (although, ideally, we'd have worldwide disarmment).

Unless somebody can make some sort of extremely convincing argument that by deploying this tech we are taking an existential risk, rather than purely banal ones, it doesn't seem reasonble to prevent deployment.