AI already has notable risks, but what are some of the risks not being discussed today?

How AI productivity gains are quietly unlocking accessibility and agency with new tools.

In training AI systems to act more like a human brain, are we introducing very humanistic maladaptations in real time?