AGI will be superhuman from the moment it arises because computers already have vastly superior memory, calculation, attention and other capabilities than humans. Thinking AGI will be "human level" for any length of time is like thinking that a swimming pool full of gasoline will reach "human temperature" for any length of time after you throw a match into it. That isn't how this is going to work at all.
Artificial super intelligence (ASI) will likely evolve very quickly, from an initially condition of being a lot smarter than the smartest human in Day 1 to being unfathomably more intelligent and effectively godlike a few weeks or months later. At that point, it's impossible to know what it will be capable of. But it is probably a safe bet that it will be able to understand biology to a level we can hardly imagine, and will likely have some amazing ideas for how to advance medicine.
For example, ASi might be able to invent actual science fiction level nanobots quite quickly, and then those nanobots could in principle help repair almost any injury or cure almost any disease imaginable. In other words, ASI might find ways to advance medical technology that don't involve anything like our current approaches with drugs, surgery, or anything like that.