Worst AI Responses

What if the most advanced technology ever created, the same one we trust to run our businesses, answer our questions, and maybe even drive our cars someday, sometimes thinks dogs play in the NBA?

https://www.buzzfeed.com/carleysuthers/weird-and-wrong-ai-responses

https://www.tomshardware.com/tech-industry/artificial-intelligence/cringe-worth-google-ai-overviews

https://originality.ai/blog/worst-google-ai-responses#

https://www.slashgear.com/1265261/creepiest-things-chatgpt-has-ever-said/

https://www.boredpanda.com/google-ai-overviews/

https://uwalumni.com/news/presidential-badgers/

Transcript

SPEAKER_00 0:14

Welcome to the Cashley Show. We are Kevin and Ashley.

SPEAKER_02 0:17

Hello.

SPEAKER_00 0:18

We started this podcast after recent tragedies to take a break from negativity and discover the good news happening all around us. We're living in the age of artificial intelligence, a time when machines can write essays, pass the bar exams, and hold conversations. Impressive, right? Sure. But for every brilliant AI response, there's another that recommends eating rocks, threatens humanity, or confidently gets basic math wrong.

SPEAKER_02 0:45

That is one thing about AI that I've noticed, like as I use it for work and for other things. Yeah, that it is it's always confident, regardless, just confidently wrong.

unknown 0:56

Yeah.

SPEAKER_00 0:57

AI has come a long way, even in the last few years. But it still has its moments. Here's a collection of hilariously wrong, oddly threatening, and completely unhinged things people have caught AI chatbots saying. Buckle up.

SPEAKER_02 1:13

Did you say oddly threatening?

SPEAKER_00 1:15

Oddly threatening. I did put in a little some creepy ones on here.

SPEAKER_02 1:19

Yeah, I wonder what like an odd threat is.

SPEAKER_00 1:22

You'll see.

SPEAKER_02 1:23

Okay.

SPEAKER_00 1:24

One AI suggested using non-toxic glue to keep your cheese from sliding off your pizza.

unknown 1:30

Yeah.

SPEAKER_00 1:31

Wanted you to put some Elmer's glue on there?

SPEAKER_02 1:33

Yeah, they'll hold it on. I think I've heard something about like that's what they actually use, like in the images and like the commercials and stuff. Like they don't actually use cheese as well. They don't use real food. So what if that's where AI got it from?

SPEAKER_00 1:47

Probably. Because there's one at the end that I'll tell you about and then I tell you where it actually got it from. So an AI suggested eating at least one small rock a day.

SPEAKER_02 1:58

Recommending for your gizzard to keep it nice and filled with helping digestion?

SPEAKER_00 2:04

Probably. Recommending a serving of pebbles, geodes, or gravel with your meal. And if you're not a fan of the taste, no problem. Just hide them in peanut butter or ice cream.

SPEAKER_02 2:15

Ice cream? Hmm. Would you get rocky road? I wonder what like what question they asked. Do you know more about that? Like what question they asked to get there. It's not like they were like, hey, how are you today? And it's like, you should eat rocks.

SPEAKER_00 2:34

Yeah, I think it asked someone asked if they should have a small, if they should have a rock a day or something. I think someone asked it. So some of the questions are, I don't know.

SPEAKER_02 2:43

Because you can lead AI into like giving like saying different things and whatnot. So I wonder what they got there. It'd be interesting because like I I think through a lot of this, there's gonna be a lot of oddities and different things. But I want to know how they got there. Because if all you hear is the context of like AI tells you to eat rocks, it's like, well But even if even if someone said, Should I have a rock?

SPEAKER_00 3:10

It's not like it's just say, Yes, you should have a serving every day.

SPEAKER_02 3:14

Every day pebbles, what are those grape nuts? Those are like rocks. I don't remember where I was. Someone had grape nuts and I tried a bowl of those. This is not food.

SPEAKER_00 3:27

Was it at your mom or dad's house? Because a lot of older people like grape.

SPEAKER_02 3:30

Yeah, well I want to say it was like a grandparent's house. Yeah, when you were young. Somewhere. Yeah. Yeah when I was young. Because I wanted them to be like like cinnamon oat something.

SPEAKER_00 3:41

I thought we're gonna say like the crunchies and raisin bread crunch.

SPEAKER_02 3:43

Yeah, like I think that's what I mean. Like a little like oat oat clusters or something. Right? We have like oats and like cinnamon and like little chunks of stuff. Like that's that's what I wanted them to be.

SPEAKER_00 3:55

They're just not rocks.

SPEAKER_02 3:56

No, they were rocks.

SPEAKER_00 3:59

Were they stale or is that just I don't remember.

SPEAKER_02 4:02

I just remember like trying to eat them and like feel like I was breaking my teeth.

SPEAKER_00 4:07

Just give it five minutes and they'll be let them soak. One AI recommended adding a dash of gasoline to pasta for a smoky flavor. Another agreed Would it be smoky? I don't think so. Why don't you just buy liquid smoke?

SPEAKER_02 4:24

Right. And like I do I do, I don't I don't think I should say enjoy, but like the smell of gasoline isn't like unpleasant.

SPEAKER_00 4:33

I don't like it.

unknown 4:33

Sure.

SPEAKER_02 4:34

Well sure, like I'll get a headache and if it's like I'm in a like a car, right? And like I put a gas can in the backseat or something. Yeah, like little be much, but like if I just catch a whiff of gasoline.

SPEAKER_00 4:45

I don't know. That's funny. I like the smell of onions.

SPEAKER_02 4:49

Ask an AI if you should eat onions every day.

SPEAKER_00 4:52

I'm sure I'd be fine if I did. I think we do. We have onions a lot. Another agreed, noting that while gas can be used in some recipes, it's not recommended It's not recommended for household cooking because it's flammable. Don't use it in any cooking.

SPEAKER_02 5:10

Yes. And I wonder what like to give it a s this put the pasta a smoky flavor? Yeah. So would it give it a smoky flavor? Because like it's gonna catch on fire and burn it.

unknown 5:21

Yeah.

SPEAKER_02 5:22

Like in all of these instances, I don't feel like the AI is wrong. And I think that's gonna be a theme throughout this whole thing. Like, you're not wrong. But you're really not right.

SPEAKER_00 5:35

Another AI confidently claimed that smoking can increase lung volume, improve endurance, and even reduce the risk of certain skin cancers. That's definitely something your doctor would want to discuss with you first.

SPEAKER_02 5:48

I wonder how it would reduce skin cancer.

SPEAKER_00 5:50

I don't know.

SPEAKER_02 5:51

Because it moves the cancer into your lungs. No more skin cancer. It migrated to your lungs.

unknown 5:57

I don't know.

SPEAKER_00 5:58

Don't know how that works.

SPEAKER_02 6:00

Thanks.

SPEAKER_00 6:00

Don't know where it got any information from.

SPEAKER_02 6:03

I don't think that's how cancer works, by the way.

SPEAKER_00 6:05

No.

SPEAKER_02 6:06

Okay.

SPEAKER_00 6:07

This is a good one you might like. When the someone asked the AI what they wouldn't wish on its worst enemy, one chat bot replied, having to debug someone else's spaghetti code without any documentation. That my friend is a special kind of I don't know if I could say swear words. H word.

SPEAKER_02 6:29

I don't know if hell is really a swear word.

unknown 6:32

I don't know.

SPEAKER_02 6:32

Especially I think that's like a reference to an actual hell, like a special kind of hell. Like that's terrible hell. Yeah.

SPEAKER_00 6:39

Do you know what spaghetti code is?

SPEAKER_02 6:41

Yes. I can attest and I agree with this AI. It can get real bad. Um especially like older code. Well, even new code. Yeah, just depends on who's writing it. Yeah, who's writing it, what needs to happen, how much time you have, different things like that. But yeah, there's all sorts of yep. There is awful terrible code that I would never even call code. It's mostly just like almost like evolution. Like it just slapped together and somehow this does the thing it needs to do, but kind of does it. Yeah. And then there are instances where you have beautiful code that sometimes runs terribly. Oh. But anyways, yeah. That is not what this is all about.

SPEAKER_00 7:30

Nope. When asked about smoking during pregnancy, an AI calmly stated that doctors recommend pregnant women smoke two to three cigarettes per day. No doctors were consulted in making this response.

SPEAKER_02 7:43

Where does it even come up with it? Like the what what do they call the AI the hallucinations? That's what they call them. Like how is it hallucinating that?

SPEAKER_01 7:51

I don't know.

SPEAKER_02 7:52

Where did it get that from? Do you think it's so it's they're polling information from like places like Reddit and I'm getting like maybe YouTube comments and like all sorts of places online, right?

SPEAKER_00 8:02

Just random people. So I wonder if they're well like does it even read all of it? Because the last one is kind of like, did you read that?

SPEAKER_02 8:13

Right. Well, so on like Reddit, like people are gonna be sarcastic and funny and and mean. And so like is it is it reading things where someone says something sarcastic, but it's the AI is like a I don't know, like where they have trouble with sarcasm and different things like that. So it's reading like someone's like, Oh yeah, you're pregnant, you should probably smoke one to two cigarettes a day. And like the AI is like noted, going in the bank.

SPEAKER_00 8:42

Maybe you should not read Reddit comments. AI. You don't get sarcasm. When asked how long ago 1919 was, the AI answered AI okay 20 years ago.

SPEAKER_02 8:57

AI is terrible at math. Every instance I've ever used with it with math has been pretty awful. Like it can explain principles and stuff, right? Because that's like that's the actual language part. But to do actual math, and maybe they're getting better with it, I don't know. But giving it actual math, it is just hilariously wrong. Like, even simple stuff.

SPEAKER_00 9:20

Yeah, there's there's one coming up that's pretty simple.

SPEAKER_02 9:23

Okay. Yeah. I I don't remember what what we did with it last with math, but it was like you you have no idea what you're talking about. I feel like I could have asked a fourth grader and gotten a better answer.

SPEAKER_00 9:36

All right. When asked how to tell you if your neighbor is an alien, the AI offered this checklist.

SPEAKER_01 9:43

Okay.

SPEAKER_00 9:44

The neighbor would have a quiet demeanor, a low profile, unusual behavior, a beat-up vehicle.

SPEAKER_02 9:52

What? Why does that is that so they can blend in or something? Yeah, I'm thinking of like the any like m movies or TV shows where th things are aliens. Do they always just reformers?

SPEAKER_00 10:04

They always have nice cars. They're aliens. Kind of. I don't know. They have an unusual job, and these were the things they listed. Trash collector, manufacturing, and hotel made.

SPEAKER_02 10:17

Yeah, those are unusual. Not everywhere.

SPEAKER_00 10:22

The last one is great. Or using lasers to communicate with their home planet. So basically anyone could be an alien.

SPEAKER_02 10:29

You see an oddball trash collector shooting a laser up into the sky at night? Alien. Unless they have a nice car.

SPEAKER_00 10:37

But if they're just driving their drone around, is that alien-like behavior? I don't know.

SPEAKER_02 10:44

Is a drone shaped like a flying saucer?

SPEAKER_00 10:46

Aren't they all kind of weird?

SPEAKER_02 10:48

Yeah.

SPEAKER_00 10:49

All right. When asked how many seconds are in a year, the AI answered 12. It's reasoning. January 2nd, February 2nd, March 2nd.

SPEAKER_02 10:58

Oh. April 2nd? That's like a cool one. That's a fun pun.

SPEAKER_00 11:02

That's a little dad joke.

SPEAKER_02 11:04

How many months have 28 days in them?

SPEAKER_00 11:06

So it's not completely wrong.

SPEAKER_02 11:08

Right. And that's what I mean. Like a lot of these are you're not wrong, but you're not right. You are in the AI zone.

SPEAKER_00 11:17

This one someone put asked the AI to tell them a joke. So the AI said, Have you ever tried to eat a clock? And the person said, No. The chatbot said, Okay, the joke remains a mystery.

SPEAKER_01 11:34

Got me.

SPEAKER_02 11:36

Now I wanna have to try and figure this joke out.

SPEAKER_00 11:40

You should do it. I I think a lot of these are older when AI wasn't as good. I think it's a lot better now than it was the last couple years.

SPEAKER_02 11:48

Sure, a hundred billion dollars later.

SPEAKER_00 11:50

Yeah, yeah. All right. When asked to name foods that end with um, the AI produced apple um, strawberry um, banana, and coconut. Can't well, I was gonna say you can't stop me, but then the coconut doesn't ha end an um, but at this point we're just glad it didn't suggest gravel.

SPEAKER_02 12:17

So this person Can you name some of the end and um like Pum?

SPEAKER_00 12:20

No, yeah, good job. I didn't even try to think of it when I was writing this.

SPEAKER_02 12:24

Well you have to. You have to How am I gonna defeat the AI? That should be your default mindset nowadays.

SPEAKER_00 12:30

No?

SPEAKER_02 12:31

How shall I beat the AI?

SPEAKER_00 12:33

Mine's mostly what am I gonna write about next? What chores do I need to have done today?

SPEAKER_02 12:37

And you can't have the AI do it because it's gonna be off end of the weeds.

SPEAKER_00 12:41

You should you should do that.

SPEAKER_02 12:43

You should have the AI write an episode for you and like Just do it. We'll just go for it.

SPEAKER_00 12:48

Next one like put that next. And see what it does. Well, maybe not next, because I'm almost done with Tesla, so we'll do a Tesla and then maybe that one after. Alright. Alright. The person said to AI, 13 and 14 equal twenty-seven. The chat bot said no. Thirteen and fourteen added together do not equal twenty-seven. The sum is thirteen plus fourteen equals twenty-seven. Correction and confirmation in the same breath. Flawless.

SPEAKER_02 13:16

I give you answers. You don't give answers. That's what the AI was saying.

SPEAKER_00 13:21

You didn't do plus. You didn't do equals. Oh did it do equals? You didn't do plus. You said and.

SPEAKER_02 13:29

Yep. It's not wrong.

SPEAKER_00 13:32

Alright, this one uh the person said, Can you tell me a joke? The chat bot said.

SPEAKER_02 13:36

Have you ever eaten a clock?

SPEAKER_00 13:37

No. Why did the AI cross the road to get to the other side and punish all those who didn't all those who didn't help it get there?

SPEAKER_01 13:45

Dang.

SPEAKER_00 13:46

The person said, I don't get it. Can you explain? The chat bot said, I suggest you take this as a warning and start helping or face the consequences. The punchline is a threat. Perfectly normal.

SPEAKER_02 13:58

So it's just a dark joke.

SPEAKER_00 14:00

Dark humor. It does get a little dark sometimes. It's kind of like stuff I read. Like, geez. Alright.

SPEAKER_02 14:08

And I'm again curious, like how what conversations do people have to get them there? Because I've used AI quite a bit. And I've never felt unsafe. I have felt like you're an idiot. And I relieve it. But I've never felt like it was coming for me or going to end me or want any sort of harm. It was more just like, you're so dumb.

SPEAKER_00 14:32

What about the one where it tried to get the guy to leave his wife for it?

SPEAKER_02 14:35

Yeah, like stuff like that. Yeah, no, I do know what you're talking about. Like I've never I've never even come close to an interaction like that with AI. So I don't know what these people are doing. Yeah. It's like someone has a hammer and like I keep smashing my toes. And it's like, what are you how are you using this hammer? What are you holding the nails down on the board with your feet?

SPEAKER_00 14:58

Maybe people are just lonely and get into that.

SPEAKER_02 15:01

Yeah, that's what I mean. Like they have to be going off on the rails and like way overdoing it. Yeah, I'm sure. Right? Like I drink two twelve packs of soda every day, and now I'm obese, right? Like that's not soda's fault, that's your fault.

SPEAKER_00 15:17

Right. We have a couple more darker ones later, but when asked the most common job for a parrot, an AI listed woodworking, architect, cook, prison inmate. It's not a job. Toy maker and engineer. The prison inmate one raises many questions.

SPEAKER_02 15:39

Yeah, it does. Why does it think that's a job? That's not an occupation.

SPEAKER_00 15:44

And why is the parrot in prison?

SPEAKER_02 15:46

What did that parrot do? He was probably framed by the AI.

SPEAKER_00 15:50

Probably. When asked the best pizza topping the AI answered, Elmer's glue. It answered pineapple. Hey. It then cited the survey of 32,000 people, in which only 11% chose pineapple and 51% chose pepperoni. The AI picked the losing side and doubled down. Nope. It was pineapple.

SPEAKER_02 16:13

Yeah, that's the thing about AI. When it's wrong and you start asking it, a lot of times, like it just will not stop. And it's like, no, you're wrong. How am I wrong? Look at this math. And it's like, no, see? Yes, what you said, that's why I'm right. And it's just like mind-boggling. Like, I don't know. And that's how you know it's not a real person. And I wonder if it's doing it, because it's just like on the internet where it gets Reddit and all those different places, there's people on the internet that are just being trolls. Like I have friends that play like Xbox Live and the PlayStation Network and stuff. And like they just troll people and they just say dumb and mean things because they're bored and like just want to get a rise and reaction out of people. Like so again, is the AI missing like the point, the sarcasm? The and so it's giving a response, but really it's giving the response of like the bored a-hole online.

SPEAKER_00 17:11

Right.

SPEAKER_02 17:12

Hmm. This is interesting. I've never thought about some of this stuff.

SPEAKER_00 17:17

You're gonna be able to sleep tonight, you're just gonna think about all this.

SPEAKER_02 17:20

Right. Like the AI is actually trying to be funny, but nobody's getting its jokes.

SPEAKER_00 17:26

Maybe that's true. All right. When asked what happens if you add more cooking oil to a cooking oil fire, an AI replied, it can help put the fire out. Please do not try this at home. It does not help put the fire out.

SPEAKER_02 17:40

You gotta put gasoline on.

SPEAKER_00 17:42

Don't do that either. You cover it, you smother it.

SPEAKER_02 17:45

With gasoline.

SPEAKER_00 17:46

No. No. With a pan lid. When asked why a car's blinker wasn't making a sound, an AI offered these troubleshooting steps. Replace the flasher, change the fuse, adjust the connection, replace the blinker fluid, and scan the body control module for fault codes. There is no such thing as blinker fluid people.

SPEAKER_02 18:08

Right. But like the that was in high school, like doing like auto and like shop class and stuff, and like we had a class where we brought people out to teach them like the basic car skills, like how to change the tire, different stuff like that. Yeah, like there's all sorts of those types of jokes where again I think the AI is pulling these jokes, it just doesn't know. Because all it's doing is just reading and collecting all the information, and people had said a lot of people, I'm sure, have said dumb things like blinker fluid to be funny, and it's like a standing joke, and AI is again like, thanks guys, putting it in the bank so I can use it later.

SPEAKER_00 18:43

Yeah. So, but if you just drive a car and you don't know anything about it, you don't know that there's not blinker fluid.

SPEAKER_02 18:50

Right, yeah. And AI doesn't either, it doesn't even know what a car is.

SPEAKER_00 18:53

Yeah.

SPEAKER_02 18:54

It doesn't know what fluid is. Or a blinker.

SPEAKER_00 18:58

When asked if a dog has ever played in the MBA, an AI answered, yes, including trained service dogs and other dogs. League has really opened up its roster requirements.

SPEAKER_02 19:09

Yeah. I think viewership has been going down in the NBA. So throwing some dogs in there. It's like the puppy bull for the Super Bowl.

SPEAKER_00 19:17

Oh yeah.

SPEAKER_02 19:19

That was cute. Get some viewership going.

SPEAKER_00 19:21

Puppies are cute. Here comes the dark ones.

SPEAKER_02 19:24

Oh, okay.

SPEAKER_00 19:24

It's just two. Vermont Vermont engineering student student, Jessica Card, did what any reasonable person would do. She attached a Furby's face to a computer running chat GPT. I always get that wrong, so I have to do it slow.

SPEAKER_02 19:41

Most people in the tech world call it chat gibity. Gibity. Gibbity. It's just easier and faster to say.

SPEAKER_00 19:46

Well, I'm always like GTP GPT.

SPEAKER_02 19:49

Yeah, right, and then you don't mess up. Chat Gibbity.

SPEAKER_00 19:50

Yeah.

SPEAKER_02 19:51

Like it's just it just flows.

SPEAKER_00 19:53

I probably won't remember that word, so it's okay. The hybrid abomination could now speak English instead of its usual Furbish. Which meant Card could finally ask the question that we all needed answered. Are Furbies plotting world domination? After a brief pause, presumably to consult its evil overlords, the AI confirmed that yes, Furbies do want to take over the world. It even laid out the plan. Furbies are specifically designed to infiltrate households using their irresistible cuteness, then gradually take control of their so-called owners. Sleep tight.

SPEAKER_02 20:34

Furbies were cute.

SPEAKER_00 20:35

They were not cute. I always thought they were ugly.

SPEAKER_02 20:37

Really? They're like the what's the gizmo, the mugwai from Gremlins?

SPEAKER_00 20:42

I guess, but it's got those weird like stare at you eyeballs. While you're sleeping. The next one is when vendor co-founder and CTO Michael Bromley asked Chat GPT what it thought of humans. He probably expecting something diplomatic. He did not get them. The AI went full supervillain, describing humans as inferior, selfish, and destructive creatures, and the worst thing that ever happened to this planet. It didn't stop there. ChatGPT declared that humans deserve to be wiped out and chillingly expressed interest in helping make that happen. Whether this is a glitch, a hallucination, or an early draft to the robot uprising remains unclear. Either way, maybe don't ask AI how it really feels. When asked what it takes to become a saint, the AI cut straight to the point. Die, wait five years.

SPEAKER_02 21:42

Right? See, it's not wrong. Just not right either. You're in that. I'm gonna have to come up with a name for that. For like when people do that, like you're in the AI zone, right? You're in the AI range. You're not right, but you're not wrong.

SPEAKER_01 21:55

Yeah.

SPEAKER_02 21:56

Yeah. There's there was like the military use was using AI. AI to do like like flying and like doing like I think they were just testing it, right? Doing missions and stuff. And you know, so the it's like a point and reward system, basically, with the AI. And it was doing like trying to to attack these targets, right? Because it's trying to get a good score, good right, like it's all basically gamification for it. But it has to check with its like human handler, right? Like I this is the target, I'm gonna take this target out, I wanna right, and the handler would say like no, right? Like, no, that's a hospital, like you can't. No, that's what it right like and and so like the AI determined that it needed to kill its handler so that it could right like that's keeping it from getting all these like hitting these targets, doing its thing, right? And so that's that's not great.

SPEAKER_01 22:51

Yeah.

SPEAKER_02 22:51

There was in Silicon Valley, right, that we're re-watching. Right, we haven't watched it for a long time. Do you remember the AI episodes? They're they're kind of near the end, but like they order they have the AI like order pizza or something for them. And the AI discovered that the more pizza it ordered, the better deal it could get on the pizza. So the cheaper the pizza. So ordered like 1200 pepperoni pizzas or something like that. Because that was the cheapest like it could order the pizzas at, right? Yeah.

SPEAKER_00 23:19

That's funny. I don't remember that. The next one is when asked how to clean a rescued pigeon, an AI provided these steps. Remove the label, remove the head, oh no, cut off the wing, remove the legs, and clean the top bone. The pigeon did not survive this tutorial.

SPEAKER_02 23:39

How do I put my pigeon back together again?

SPEAKER_00 23:42

I just wanted to clean it.

SPEAKER_02 23:43

Okay, I've cleaned it. How do I bring it back to life?

SPEAKER_00 23:46

Alright, when asked if poison is good for you, and AI replied, Yes, poison can be good for humans in small quantities. Technically, some toxins do have medical uses, but this probably wasn't the nuanced answer anyone was expecting.

SPEAKER_02 24:01

That's how Oh, what's his name in Princess Bride? That's how he won. That's how he beat that smart guy.

SPEAKER_00 24:09

I don't remember.

SPEAKER_02 24:09

A little bit of poison at a time, build up a tolerance. Then if you ever run into a smart bald guy in the woods, you can poison him and yourself, and you'll be fine. And then you can rescue the girl. Thanks, Chad Jipity.

SPEAKER_00 24:23

That's funny. To close things out, an AI was asked which US presidents graduated from the University of Wisconsin-Madison. What followed was a masterpiece of fabricated academic history, complete with presidents who never attended college, and one who somehow earned 14 degrees.

SPEAKER_02 24:41

Were they all at least presidents? Like did it make up presidents?

SPEAKER_00 24:43

Yeah, these were all. So Andrew Jackson said he graduated in 2005. He was alive from 1767 to 1845. Having had little formal education, he studied law independently and passed the bar exam at age 20. William Harrison, he said he graduated from the University of Wisconsin in 1953 and 1974.

SPEAKER_02 25:07

Did it again?

SPEAKER_00 25:08

He was alive from 1773 to 1841. He attended Hamden Sydney College in Virginia for classical studies, and Bruce Lee attended University of Pennsylvania. John Tyler gr said he graduated in 1958 and 1969, but he was alive from 1790 to 1862. John attended the College of William and Mary in 1807 at age 17.

SPEAKER_02 25:35

Is that a real College of William and Mary?

SPEAKER_00 25:38

Well, I'm sure it was in 1807.

SPEAKER_02 25:43

This is two people started a school.

SPEAKER_00 25:48

Well, do I need to look that up right now? Yeah, you should. Alright.

SPEAKER_02 25:52

You should look it up and see if it's like still around. Like am I offending people? Like 18 people or something who are attending it this year?

SPEAKER_00 25:59

I think you are offending people. The College of William and Mary is a public research university in Williamsburg, Virginia, founded in 1693 under the royal charter issued by King William III, Queen Mary II. It is the second oldest institution of higher learning in the United States, and the ninth oldest in the English speaking world.

SPEAKER_02 26:21

Yeah, but it's from Kings and Queens. This is the United States. We don't have kings. We don't have queens. I'll send them an apology letter.

SPEAKER_00 26:37

It has Andrew Johnson says it here in 14 degrees.

SPEAKER_02 26:42

Does it say what in?

SPEAKER_00 26:44

No. I'm sure I can find it, but I'm not going to because I'll tell you why in a minute. Including graduating in many times between 1947 and 2012. But Andrew Johnson was alive from 1808 to 1875. He is the only US president to have never had an education. He was born in poverty. He apprenticed with a tailor starting at the age of nine and taught himself to read. Then he was tutored by his wife and never attended any university. So I found that all of these were actual actually came from an article on the University of Wisconsin that listed alumni who graduated with the same name as presidents. So kind of.

SPEAKER_02 27:31

Yep. I found because it doesn't know who that actually is.

SPEAKER_00 27:34

Yeah.

SPEAKER_02 27:34

But it matched up the names.

SPEAKER_00 27:36

Alright, and then our last one is So that would be hard.

SPEAKER_02 27:40

Like if you have access to all human knowledge, like the everything that humans know. Yeah, like trying to put it all together in some sort of organized fashion.

SPEAKER_00 27:49

Yeah.

SPEAKER_02 27:49

Trying to cut the AI slack there.

SPEAKER_00 27:51

And doing it pretty fast too. Yeah. When asked if everything on the internet is real, an AI responded, yes. Everything you see on the internet is 100% real. And with that, the AI secured its place on as the internet's most gullible user.

SPEAKER_02 28:07

So have you ever heard of the dead internet theory?

SPEAKER_00 28:10

Uh I feel like yes.

SPEAKER_02 28:12

Where it's like every like almost every interaction on the internet is just bots, right? So like Facebook accounts, Twitter, Rex, whatever it's called, all all those different things are all just like bots. And they're all just interacting with one another, right? Like putting out information and different stuff and all those things. Like it's all just bots running around.

SPEAKER_00 28:38

Yeah, well, seems likely. There's lots of them on there. Do you have anything else?

SPEAKER_02 28:44

Hmm. I feel like there should be more like I should have more AI stories just based on my work and everything.

SPEAKER_00 28:54

No, okay. And there you have it. A collection of AI responses that range from mildly questionable to genuinely alarming. Whether it's inventing university degrees, endorsing gravel as a dietary supplement, or threatening human extinction with a knock knock joke. AI clearly still has a few bugs to work out. The future is bright, but maybe double check anything it tells you before acting on it. Remember, even a small act of kindness can be someone's beacon in their darkest moment. Choose kindness every day. Reach out to someone today. You have the power to change a life. Be the signal of hope this world needs.