Can AI catch its own kind? How to check if your AI content passes the human test?

My wife was eager to test today's featured tool!
With AI-generated content becoming ubiquitous online, schools and businesses alike are scrambling to implement detection tools. One player, undetectable.ai, claims to assess content for AI origin and then, controversially, "humanize" it to evade detection.
The jury's still out on its claims. My own testing puts its accuracy at classifying AI content at around 65%, and while it does decent with the "humanizing" bit, it's far from perfect just yet.
Ese