This is a perfect example where AI is used / working alongside humans to get the best of both worlds. Humans attempted to fix it and failed multiple times before. Now AI has a fix, which is still not perfect, but still promising. Now, on the current state with the known issue, both humans and AI can look for a fix.

Bit by bit, they will together hopefully come to a solution where it both works, and everyone agrees it works and nobody finds any issues.

Was that so hard?