There is no such thing as 100 percent, and that is also applicable to all Artificial Intelligence (AI) detectors. For instance, I tried to write a recent blog. Every time I finished my draft, one of the things that I usually do is to use different AI detectors to test if it will detect my work as an original content, or it will be flagged as AI generated content. Upon doing so, there is a small percentage of my content that has been flagged. But the truth is I created the whole thing myself. So, where's the issue lies?
In that case, we can say that even though AI detectors comes with thorough programming logics, deep training, and machine learning, it will always be flawed.
The thing with other people who publish their works here on Hive is that they often rely on AI, to the point that the human essence fades. For the sake of rewards and easy published contents.