Ai can accelerate a lot of things but drug development is mostly depending on long processes involving human reviews and human trials that can't be shortened unless you do a massive regulation reform.
There's likely lots still to discover about biology and A.I. doesn't generate new data, rather rearranging old data. That's still helpful but developing new tools to make basic science discoveries to expand the knowledge pool seems to be the rate limiting factor rather than rearranging existing knowledge.