SIA Toolkit26 Lez. 1.3 – AI Detector: puoi davvero fidarti? False e accuse e funzionamento reale.

Here we are. About two weeks ago I posted a video about the Falad case, the author who apparently lost a contract worth $2 million or more because there was a suspicion, and the impossibility of refuting it with a documented reconstruction, that his novel had been written by artificial intelligence. That video caused quite a stir on the

Rotte Narrative YouTube channel, and something happened underneath that video that gave me even more food for thought than the case itself, and now I want to show it to you. There you have it, there were comments both under that video and under the other videos that it paved the way for. For example, there was

Emy's comment which I'll summarize, I won't quote it word for word. I tested my first novel, written in 2010, on AI-debunking sites. Uh, 70%. [laughter] Eh, if not, Erica another commenter, if you give them a full AI text, sometimes they say it's not. And a third comment, I think in a different video.

Marco wrote text eh written by two hands the detector le Ai detector gives result 65. After my substantial changes 100%, 100% of ya. [laughter] They are completely unreliable. Here, then, are three normal people, three individual experiences that, individually, would be anecdotal, but taken together, make it clear that the overall picture—the overall picture—is

something amiss. So, how much truth is there and how much relevance is there now that artificial intelligence has become more advanced , is the idea of ​​being able to identify in a text through an AI detector the parts that were written by artificial intelligence. Well, this is the question I asked myself . I said, I should go