Yes, “Chat” AI is biased. But…
I have spent most of my working life as a Software Engineer. However, I spent most of that time as a database developer so, though I thought it would be fascinating, I never worked on anything even close to Artificial Intelligence.
Following with interest how others have responded to the various publicly available “chat” AI interfaces, I finally broke down and connected. (I can’t put my finger on why, exactly, but I always felt intimidated by the prospect of speaking directly to an AI. One of those little pieces of irrationality that seems to have sneaked into my psyche — which I have already begun the process of sweeping away.)
In reading the questions others had asked and the responses they got, I realized that “chat” AI is mostly an interface between the user and a good Internet search engine. So I had to ask something that would help me isolate as much as possible the interface part from the search engine part.
This consists of testing how the AI interpreted my questions and how it interpreted the data it got from the search engine. I was also aware of several charges that the “chat” AIs had a political bias designed into them; so I wanted to test for that, also.
