Google's Experimental Bard AI Has the "Large Libel Model" Problem that ChatGPT-4 Has
I blogged yesterday about how ChatGPT-4 produces output that contains false allegations about particular people—falsehoods apparently created by its own algorithms—and attributes them to prominent media outlets. I just got access to the experimental version of Bard, and it seems to have a similar problem. It's a little less quick…
Reason ·