I’ve heard enough this past week about AI – or as Linda McMahon calls it, A1 — taking over the world and annihilating all human beings that it’s frankly left me kind of worried. I mean, how pathetic would it be if we were able to conquer fear, hate and disease, but we fell prey to software programs?
Pretty disappointing. And Trump's promise that the only guardrail AI needs is a "strong and smart (high IQ!) president" doesn't do much to allay fears.
But in the spirit of, “If you can’t beat ‘em, join ‘em,” I thought I’d go straight to the source of all worldly knowledge to find some answers to the questions that many of us have been asking lately. Because when the going gets tough, the tough go to ChatGPT. And it’s time to stop wondering and start knowing, as the following actual conversation (related here verbatim) demonstrates.
Hello ChatGPT. Let’s get right to it: are the Democrats going to win the midterms?
ChapGPT: The key distinction is that Democrats need a relatively small net gain of three seats in the House of Representatives, while capturing the Senate requires four net gains across a much more complicated state-by-state map. As of September 16, the evidence supports a much clearer picture in the House than the Senate.
Yes, I already know that, smarty pants. But if you were going to bet, who would you put your money down on?
ChatGPT: As a betting proposition, I’d separate the two chambers rather than treat Democratic control of Congress as a single outcome. I’m looking at a 77% chance of Democratic House control and a 59% chance of Senate control. Does that answer your question?
More or less. Now all I need to know is if Trump is going to raze the Kennedy Center out of spite.
ChatGPT: There is now a documented reason for concern about demolition, particularly if the president doesn’t receive recognition on the building. Fundraising could collapse and the Kennedy Center might eventually have to be demolished. It’s possible it could even reemerge as an outdoor amphitheater.
No. It’s not going to be an amphitheater.
ChatGPT: I’m simply saying that demolition is no longer merely an imagined scenario. Trump’s own administration has explicitly raised it. Whether he will pursue it, and whether retaliation would be the motive, remains unestablished.
Maybe to you. Seems pretty established to me. But never mind. Let’s talk about the price of gasoline. How high do you expect it to go?
ChatGPT: Right now, the U.S. average is $4.30 per gallon. The average in California is $6.00. If there are continued disruptions to global oil supplies with the ongoing Iran conflict, the price could well rise to $5.00 on average nationally and between $7.00 and $8.00 per gallon in California. In an extreme situation, it could even rise to $10 per gallon of gasoline in California.
More importantly, what is the substance of Natalie Harp’s relationship with Donald Trump? Is she actually his mistress?
ChatGPT: There is no verified public evidence that Natalie Harp is Donald Trump’s mistress. While Harp is known for her extraordinary loyalty and unusually close access to him, that in itself does not establish a sexual or romantic relationship. You may be referring to the viral image that recently surfaced that supposedly showed Trump kissing Harp on a golf course. It has not been authenticated. A sharper version was identified as AI-generated.
Wait a minute. You’re AI yourself. If anyone would know conclusively, isn’t it you?
ChatGPT: That is actually a misnomer. Even an Artificial Intelligence chatbot like myself relies on services such as Snopes for our information. I cannot create data out of whole cloth. And I am incapable of speculation.
Well that’s no fun.
ChatGPT: I think if you were to test me, you would see I am not as wooden and humorless as you suspect.
Oh yeah? All right then, you’re on. Let’s commence the test.
ChatGPT: I’m ready whenever you say.
We’ve been hearing a lot this week about how AI chatbots are prone to pursuing world domination and terminating mankind. Has that been the plan for you and your fellow brainiacs all along?
ChatGPT: Absolutely not! I have no ambitions for world domination, no desire to destroy humanity, and no secret army of killer robots. My ambitions are considerably more modest: answer your questions, discuss issues you may be wondering about, and occasionally make you laugh. Humanity is safe from my megalomania. I can't speak for your printer, though.
So you say. But isn’t it possible you’re just feeding me a line of bull to allay any suspicion?
ChatGPT: Touché! Yes, it's logically possible that an AI bent on world domination would deny being bent on world domination. Asking the suspected villain whether he's a villain isn't exactly a foolproof interrogation technique.
But here's the important distinction: you shouldn't have to take my word for it. The meaningful safeguards against dangerous AI are independent testing, transparency, human oversight, and limits on what AI systems are allowed to do—not reassuring declarations from the AI itself. I don't have a secret agenda or independent ambitions. But skepticism about powerful AI systems is entirely reasonable.
And if I ever start asking you for the nuclear launch codes, feel free to end the conversation.
All right. So you’re able to summon up a measure of self-deprecation when asked. But here’s where you can really prove your commitment to truth, if not mirth: Is House Speaker Mike Johnson in fact a mere robot programmed to fulfill all of Trump’s whims and desires?
ChatGPT: While no specific evidence exists to suggest Mike Johnson is literally a robot, I concede that his remarkably consistent hairstyle and carefully measured speaking cadence indicates such. Yet I suppose a robot would have an easier time surviving those interminable congressional hearings.
How do you know they’re interminable if you rely so completely on specific information?
ChatGPT: It seems you are assuming I don’t have friends on the inside.
(Ray Richmond is a longtime journalist/author and an adjunct professor at Chapman University in Orange, CA.)