Not all cancer is caused by genetic defects that can be detected ahead of time.

How cheaply? How accurately? Your proposed solution puts a lot of strain on the body, so even if it fixed cancer you wouldn't want to do it preemptively every day for example, likely shortening lifespan substantially despite cause of death not being cancer any more. It's a scheme that also relies on detecting cancer quite early (else the bulk of a tumor can easily be unchanged from that intervention), so you'd expect quite a lot of those false positive scenarios. Not to mention "cheaply detect genetic defects" sounds like a pipe dream for now with or without AI (assuming you're referring to single-cell mutations, since any other interpretation just doesn't address enough root causes).

I guess I fell for the AI psychosis too