[raaman]

I just checked one paragraph of an article I wrote here in 2016 on an AI content detector tool. It shows the result as 84% AI-written content. How is it possible? In 2016, AI tools were not at all invented and were not in existence.

[Netherrealmer]

One thing I notice about AI is it loves using the following words. Captivating. Wild ride. Rollercoaster. Eager. Eagerly. And well, well, well. Also, check the country of the writer. Even if a person is fluent in English, a person of a certain country has their own way of speaking English. And if they use words most people from that region are not using, and you will check their past posts and creations. It's easy to detect. For example, I notice a lot of African members use AI a lot. I know how Africans speak English. So I can detect if they're using AI. Because Africans have their special way of speaking English. For example, most Africans call TV show a movie. Not all, but most of the time.

Out of nowhere that person speaks victorian english in just few days.

Heck, even Native English speakers have their own way of speaking. And if they suddenly speak out of nowhere like how they used to speak few days ago, most likely they're using AI.

[germainebull]

It's really amazing! 🤔 It's possible that your text element has a style similar to what AI uses today, so the detector sees it as being made by AI.

Also, some AI content detection tools are not entirely accurate. 🔍 They rely on language and word structure, not writing history. So, such errors occur frequently.