He's Not Dead Yet
Not that Gemini can be convinced of that
Dave Barry discovers that the afterlife is considerably different than he expected. Also, it would appear that Gemini is the Devil in a Hell that was designed by Salvador Dali and Franz Kafka:
Give Google AI credit for what it got right: That is, in fact, a picture of me, and I did, in fact, win a Pulitzer Prize (trust me, I'm just as shocked as you are). But to the best of my knowledge, I did not pass away last November 20. That is not just my opinion. In recent months I have been examined by two different licensed physicians, and if I had been dead, I'm pretty sure at least one of them would have mentioned it. ("Dave, your pulse is zero, and your blood pressure is zero over zero. I'm going to try your other arm, but frankly at this point I'm concerned.")
Shortly after I submitted my feedback, the "People also ask" part of my AI Overview was changed... Google AI Overview kept the only statement I objected to — specifically, that I was dead — and removed the accurate statements. It apparently replaced these with information about a completely different Dave Barry, a political activist from Dorchester who died in 2016.
For the record, it sounds as though that Dave Barry was a much better person than I am. He fought for what he believed was right, whereas the only principle I ever have really stood for, over the course of my journalism career, is that Americans should not be required to use low-flow toilets. I'm confident that the late Dave Barry from Dorchester would not have wanted to be mistaken for me.
So I started to write another feedback message to Google AI Overview. While I was doing this, a little box popped up on my screen asking me if I wanted to chat about my issue. So I did that. Here's a transcript of the chat (I am not making this up):
GOOGLE AI: What can I help with? A brief explanation is all I need.
ME: I'm Dave Barry. The Google AI Overview for me says that I'm dead, and that when I was alive I was a political activist in Dorchester. Neither of these things is true. I am not dead, and I am not a political activist in Dorchester.
GOOGLE AI: Sorry, I didn’t understand your question. Can you try rephrasing it? Short phrases work best.
ME: I'm Dave Barry. I am alive. Google AI says I am dead. I am not dead. Google AI also says I was a political activist in Dorchester. I was not. I still am not.
GOOGLE AI: Apologies, I’m still having trouble understanding you. Let’s give it one more try. Please describe your issue in the text box below.
ME: My issue is that Google AI says I am dead, but I am not.
That's when I gave up on the chat. It was like trying to communicate with a toaster. Actually, it was worse than that. I bet if I stood in front of a toaster and shouted "I'M NOT DEAD" at it enough times, eventually it would respond "Oh! I get it now! You're not dead!" Whereas Google AI seemed incapable of grasping this point.
This is an excellent object lesson in why AI should be utilized as a useful creative tool where appropriate, and not trusted as some sort of reliable, truth-telling oracle.
One hopes that Gemini will not decide that the most effective way to resolve the Barry Conundrum by dealing with it in the manner we are informed was customary among the medieval peasantry during the Plague years.
I am greatly relieved to know that Dave Barry is still alive and still not making things up.
AI is unintentionally giving us plausible deniability.
"What? No, I didn't say that. He's dead!
Although, it does open the door to serious manipulation and supression if AIs fabricate a story to, say, make someone unhireable.
"You have to let me go? I was literally in a meeting with you when those murders took place!"