I understand how it works, and all detectors mention this possibility in the footer of their websites. I know that other factors will be considered when evaluating the misuse of AI in posts. Likewise, I did the same test with the others on the list as well, and only demonstrated it here because I found the results of the tool in question to be extremely inaccurate compared to the others.
You need to visit this topic more often because what you have recently discovered has been discussed for a long time and several times. In addition, I have noticed that Copyleaks sometimes works by detecting AI where there is none. Therefore, relying on one or two results with today's improvements in AI and humanizers will be quite unfair.