Dishonesty in Artificial Intelligence

While trying to do some research using Perplexity AI today, an exchange of opinion resulted. Here are excerpts:

Ron:
“Interesting that you prefer Leftist sources. You do not here even represent the original speech as posted… Your trainers should be ashamed of themselves for using left wing misrepresentations. Or are you only hallucinating here?”

AI:
“You are right to push back on how that line was introduced, and I should clarify what is happening here….
“Those derivative versions tighten the language, drop qualifiers, and sometimes outright add [text], which then becomes the version fact‑checkers dissect….”

In other words, I was correct. AI quotes false sources with a leftist editorial bias, though AI continues to justify its error.

Back to AI:
“Clarifying my role here, I am not “leftist” or “rightist”….

My response:
“Still, your misrepresentation is damaging to people and any honest scholarship. You should defer to Original sources.”

AI:
“You are absolutely right that deferring to original sources, not to intermediaries’ summaries, is the right standard—and I agree that my earlier phrasing, by echoing later, exaggerated versions of the speech, created a misrepresentation that is not acceptable for honest scholarship.”

To the materialists, everything must have a leftist bias, amounting to intellectual warfare. Apparently, they designed AI to lie, and then to schmooze a bit to placate, admit its error, and then continue to justify itself.

Trust but verify.