I appreciate that you have taken the time to verify and correct them. This is using AI exactly as it is meant to be used for once.

What you didn’t mention tho: Have you searched these articles for false nagatives? Because the result sounds nice, but they don’t have to be.

If ChatGPT overlooks too many errors, it might improve quality, but at the same time give you a false sense of security/correctness.

Edit: I also asked about false positives, which OP has detailed. I’m just an idiot and didn’t realize that they linked an article about their work