3 ms·
Google cautions against "hallucinating" chatbots
- mdp2021 4y agoThe original (which CNA reports about) should be Maschine liefert überzeugende, aber komplett frei erfundene Antwort at https://www.welt.de/wirtschaft/article243709293/KI-Google-warnt-vor-unbedachtem-Einsatz.html https://www.welt.de/wirtschaft/article243709293/KI-Google-wa... but it does not seem to contain much more content. > "This kind of artificial intelligence we're talking about right now can sometimes lead to something we call hallucination," Prabhakar Raghavan, ... Google, ... told Germany's Welt [...] One of the fundamental tasks, he added, was keeping this to a minimum The big question is how will you minimize hallucination and how you will implement "reliability" in the old architecture.