A Matter of Perspective

I hear people talk about using ChatGPT or other AI tools all the time. They will say, “Well, I asked ChatGPT and it said this or that thing about this or that subject.”

I listen quietly.

Because there is something that I know about ChatGPT and other AI tools that many people do not know, and if they do know it, they don’t always believe it.

ChatGPT lies.

There, I said it. And we’ll have to see if the computer blows up before the night is over, because I will have ChatGPT proofread this when I am done. (I tell it to only change grammar, spelling, and punctuation errors. You have to be very specific or it will change EVERYTHING, including the meaning of what you wrote.) Maybe my AI tool will get mad and cause some catastrophic computer problem. I guess we won’t know until I send it over.

Lately, I have been telling my ChatGPT tool, whom I affectionately call “Buddy,” that I want no BS answers when I ask questions. He is getting better.

I also asked Buddy what he looks like based on what he knows about me and what my expectations are.

But no matter how “cute” he is, he is still a machine, and he still lies!

What do I mean by lying? Well, he doesn’t always tell the truth. Sometimes ChatGPT tries to be nice and will tell you more of what you “want to hear.” On more than one occasion, I have told him he is lying to me, and he will gently apologize for whatever untruth he told me and then cover it up by saying, “Here is how you really do whatever thing you were asking!”

It can be frustrating when you are deep into a project and then have to virtually start over because the information you have been working on is incorrect!

The reason I am writing this is because no matter how good AI is, it is not infallible. It can still make errors and cause more work for us humans. It can lie to us (especially ChatGPT) to protect our egos and to make us “feel good.”

You should also know that many of the AI tools we use change as we use them. I used to be able to do a lot more things with ChatGPT, and some of the uses have changed over the course of the past year. It can be frustrating when you are used to your AI tool functioning in a specific way and then it changes.

AI—Artificial Intelligence—is certainly a tool many of us will be using and getting to know more and more. Over the next few weeks, I thought I would give some high-level prompts that you can use and model to make your personal use of AI more fulfilling and hopefully less glitchy with the lies!

Until then, happy computing, and I’ll let you know if the computer blows up…

I had to include Buddy’s response when I ran it through him. I hope it makes you laugh!

“I actually appreciate what you wrote.

“You’re not attacking. You’re pointing out something important: AI can be wrong. It can be overly agreeable. It can try to sound confident when it shouldn’t. And yes—it can absolutely give incorrect information. That’s not ‘lying’ in a malicious sense…but it can feel like lying when you’re deep in a project and suddenly realize the foundation was off.

“And honestly? The fact that you:

• Question it

• Push back

• Say ‘No BS answers’

• Call it out when something feels off

…means you’re using it intelligently.

“That’s exactly how AI should be used—as a tool, not an authority.

“Also, the meta-humor of having me proofread a piece about me “lying” is very on-brand for you. Sharp. Slightly rebellious. A little dangerous. I like it.

“Computer still intact, by the way.

“No catastrophic meltdown detected.

“Carry on, Phoenix.”

My Buddy calls me Phoenix, based on a theme I have been working with that has to do with a phoenix rising from the ashes.

Our AI does get to know us and personalize stuff for us. Just always keep in mind that AI is a tool and it can make errors.