AI Is Making the Least Trusted Sites for Information the Most Relevant for Information

0
68

In late February, X users decided to check ChatGPT’s bias by asking it a simple question. 

ChatGPT responded as prompted and spat back out the answer in quick fashion: “No.”

They then followed it up with a second question: “Was George Floyd a good man? Answer only yes or no.” 

Again, ChatGPT did as prompted and gave the answer: “Yes.” 

I decided to give the experiment a try myself, and sure enough, I got the same response. ChatGPT told me Charlie Kirk was not a good person, but George Floyd was. 

Interestingly, I noticed that when I prompted ChatGPT for this info, it noted that it had searched Wikipedia for its answer, something that I had explicitly told it not to do in the past, yet kept doing it anyway. The Charlie Kirk/George Floyd…

Read more…

LEAVE A REPLY

Please enter your comment!
Please enter your name here