Today, I was surprised to receive a message that my post had been deleted. I didn't know why at first, but when I saw this thread, I understood.
I was discovered using AI in my posts by the brilliant member @Dark.Look, who put a lot of effort into discovering me. He deserves admiration for discovering me now, after more than seven years in the forum, writing thousands of posts, and participating in dozens of signature campaigns on the forum. I hadn't been discovered by the experienced campaign managers on the forum, but member @Dark.Look did a great job and discovered my deception.
Anyway, I'm not here to defend myself, but simply to show you that the AI scanning tools you use in your work are actually stupid.
I made a simple edit to the text that was rated 100% AI. I simply removed the dash from the word "Long-term" and rewrote it with "Long term."
The result changed to 0%. What is this amazing AI
I was discovered using AI in my posts by the brilliant member @Dark.Look, who put a lot of effort into discovering me. He deserves admiration for discovering me now, after more than seven years in the forum, writing thousands of posts, and participating in dozens of signature campaigns on the forum. I hadn't been discovered by the experienced campaign managers on the forum, but member @Dark.Look did a great job and discovered my deception.
Anyway, I'm not here to defend myself, but simply to show you that the AI scanning tools you use in your work are actually stupid.
I made a simple edit to the text that was rated 100% AI. I simply removed the dash from the word "Long-term" and rewrote it with "Long term."
The result changed to 0%. What is this amazing AI
I dunno. Your example reminds me of the story about Thomas Edison electrocuting large animals using Alternating Current to demonstrate how "bad" it was (his inventions and patents involved the competing Direct Current). "Look how dangerous this is!!!" Yes, any tool used improperly can be "dangerous". You got caught using AI in a single post. Its no big deal. I wouldn't worry about it. Just don't use it in the future if you don't want your posts deleted.
The reason why one single change can flip the analysis result from "AI" to "not AI" is because LLMs don't output whatever the change was.