An interesting phenomenon. At least when AI was in its infancy, it told users what they wanted to hear. It’s getting better, me-thinks.

My evidence as to why we know AI told people what it wanted to hear is the South Park episode about it. (Peer review it, cowards.)

Still, my example as to why it’s improving is a single data point. The pessimist would call it “anecdotal evidence.” The optimist would call it “a case study.” Well . . . here it is.

I wondered how legitimate the claim was that the block chain and AI were stressing the world’s resources. I’ve taken a new strategy with AI to test it. I’ll ask the form of the question leading to the answer I want to hear. I’ll ask the form of the question leading to the answer opposite of what I want to hear. Finally, I’ll ask the form of the question that is my closest approximation to an open-ended, neutral question.

So here’s the case study. (Or anecdotal evidence if you’re an ass.)

The Question Designed For Me to Get the Answer I Don’t Want to Hear

What is the theory behind AI depleting resources?

–Google AI’s response?

The Question Designed For Me to Get the Answer I Wanted to Hear

AI depleting resources is fake, right?

And The Question Designed to Be a Neutral Inquiry

Is AI depleting resources?

So . . . my conclusion is that AI is getting more objective. Though we should ask it in three different ways to be sure.