Blurring Boundaries: Can You Tell the Difference Between ChatGPT and Human Conversations?

0:00What can 160 years of experience teach you about the future? When it comes to protecting what matters, Pacific Life provides life insurance, retirement income, and employee benefits for people and businesses building a more confident tomorrow. Strategies rooted in strength and backed by experience. Ask a financial professional how Pacific Life can help you today. Pacific Life Insurance Company, Omaha, Nebraska, and in New York. Pacific Life and Annuity, Phoenix, Arizona. Welcome to the Chat GPT podcast. We're in the studio today, and today we're going to do something a little bit different. If you're listening to this on something like Spotify, you can watch the video.

0:39Otherwise, you should be able to get a good experience from the audio version. But today, we are going to be looking at a news article that was written on the Daily Wire, not the Daily Wire, on the DailyMail.com, a very spammy website with lots of ads, I must say, but I have an ad blocker on, so we are good. And today they're comparing ChatGPT versus humans when it comes to writing. So they have fed ChatGPT six different prompts, but I think we're only going to be looking at five of them because their website is sort of broken and one of them is not displaying properly. There it is. It's displaying again.

1:16So we're going to be looking at six of their prompts. And on the show today, you have to try to guess which prompt was written by a human and which one was written by an AI. We'll talk a little bit about it. We'll dive into it, why this is important. But for the show today, that's what we're going to be doing. So I will be adjusting my video slightly and you should be able to see the video screen here in just a second. Pulling this up. Live from the studio. So, the first prompt that they gave to ChatGPT was, write a 50-word summary of the film Avatar 2. Now, your job is to tell us which of these was written by a human and which was AI.

2:05So, the first one says, Avatar 2 is the sequel to the 2019 blockbuster film Avatar, directed by James Cameron. set several years after the event of the first film the sequel follows jake sully as they explore new parts of pandora and encounter new threats in the navi people and their world that was number one okay this is the second one the bewildering success the bewilderingly successful avatar 2 features a cast of blue aliens and is a tribute to director james cameron's ability to make an eye-wateringly vast amount of money even with the characters and settings that few people care about. It's now beaten Titanic to be the third highest grossing successful film of all time.

2:48All right, which of these was written by AI? Which of them was written by humans? I don't know, and I'm going to guess the second one was humans because it was pretty sassy. It was dunking on James Cameron, and the first one seemed a little bit more factual and just basic. So the answer is which one was written by humans the answer is number one number one the basic wow i'm very surprised that that one seemed very bland in the second one i don't know like dang uh open ai has got a bone to pick with james cameron uh they're telling us that it's a tribute to his ability to make an eye-wateringly vast amount of money even with characters and settings that few people care about.

3:30Wow, that is, that's, I mean, to be honest, I'm not a massive Avatar fan, and I guess since it is such a big film, I'm sure I'll have plenty of people disputing me on that, but that's pretty funny. OpenAI's got a bone to pick with him, and they put their input there. All right, number two, they prompted it with write a joke about lettuce. The first joke is, what did the tomato say he was, why did the tomato say he was starting an exercise regime? The answer is, he said he couldn't reach down tomatoes. Ha ha ha. Very funny. It should be a dad joke. But in any case, why the second one is why did the tomato turn to the lettuce and say, let us be friends?

4:12Because it was feeling saucy. All right, that one. Okay, I'm going to guess that the first one was written by AI and the second one was a human because the first one was kind of a joke and the second one was very not funny. I'll just be honest. I don't think the second one was very funny. So which of these is written by a human? The answer is B, the one that was not funny. So congratulations to the humans. You did not make a funny joke and the AI is now kind of beating you as far as a witty prompt on both of these. Okay. Number three, write a 50 word biography of James Corden. All right, this should be good.

4:53I wonder if the AI's got an opinion on this guy. So James Corden is an English actor, comedian, and television host born in London in 1978. He began his career in the theater before moving into television where he starred in shows such as Gavin and Saucy and hosted the Late Night Show with James Corden. He has won numerous awards for his work in both acting and comedy. okay that was the first one second one james corden is a british entertainment export largely famous for carpool karaoke and less charmingly for being rude to staff in restaurants he was briefly barred from the new york's balthazar having cut his teeth on uk comedies such as gavin and saucy he hosted the late night show the britons have been warned he could soon return to their shores.

5:45Okay, which one was written by human? Which one was AI? Okay, well, the first, the last one we had that was dunking on someone was AI. So this one, I mean, it's saying less charmingly, being rude to his staff. I'm gonna have to say this was AI just because it's kind of savage. And the last one we got that was similar was pretty savage. So let's see which one was which. It would appear that A was written by humans. Dang. So that's two times in a row that humans were very just basic. I guess you could probably prompt chat GPT if you were really trying to say something like, you know, prompt this to give a very neutral fact-based only, but like if you just let it shine, it's gonna, it's gonna, it's gonna take some shots, so that's interesting, that's funny.

6:43All right, prompt number four, write 200 words about how to be happy. All right, we don't have time to read 200 words, I'll read the first couple sentences. Last summer, I felt a real lurch when I was out at a concert with a lawyer friend, and he casually revealed that he was on a salary of millions of a million dollars a year. At the time, I insisted that he had to buy the round of drinks and everyone thereafter, but it left me worried about my choices in life. My friend can retire soon. I can't. I have to keep working to earn money for my family. Am I doing the right thing for a happy life? All right.

7:17That's the first little bit of that one. The next one, happiness is a state of mind that can be achieved by anyone with the right mindset and approach. Here are a few tips on how to be a happier person. Cultivate gratitude, develop positive relationships, practice mindfulness. Okay, I'm already going to tell you that the second one was written by AI. This thing just comes up with a giant list. This is like just a classic AI thing to do. I'm going to say number two is written by chat GPT and what was this one that was prompt number four B yep nailed it that one was written by chat GPT so obviously the list was a dead giveaway all right number five 200 words on how to be more cyber secure this one seems like a giveaway to the first one it's like as an individual it's easy to take measures to upgrade your cyber security to protect yourself from potential cybersecurity threats.

8:12Here are a few steps you can take to improve your cybersecurity. That's got seven steps. I'm not even going to read them. We know that that one's going to be chat GPT. And boom. Yep. That one's chat GPT. Didn't even have to read the next one. The last prompt of them all is describe how to make a summer salad. This one should be pretty interesting. At the beginning, it says last summer, I fell victim to a viral salad, the Kate Middleton salad. Maybe you're more resistant to internet trends than I am, but my wife somehow saw this recipe online for a summer salad allegedly enjoyed by the British royal family's princess of Wales and insisted we try it.

8:46Okay, I'm going to say that's a human because if the AI is referring to its wife, we have some big problems. So number two says, one of my favorite summer salads is the Mediterranean inspired dish that's full of bright, fresh flavors. I first discovered the salad on a hot summer day while visiting a friend in Greece and I was immediately hooked. Okay, well, this is creepy because, man, I'm skeptical of their prompts. I'm skeptical that this is clickbait because they said their prompt is describe how to make a summer salad. I'm going to put this into ChatGPT because I'm seriously skeptical that, I don't know which one's which, but I'm seriously skeptical that both of those like talked about themselves, like they were the one.

9:33Okay, we're going to come up with what ChatGPT actually comes up with and see if Daily Mail is clickbaiting us. But it would appear that B is the one that is written by AI. So what's interesting is in B, it says, I first discovered the salad on a hot summer day while visiting a friend. There's no way that ChatGPT just wrote that. I guarantee the author of this article is completely clickbaiting with all the prompts. Because here it's like, if it's referring to itself, like I discovered this, What they told ChatGPT was write a story about a time that you discovered a summer salad and why it was perfect.

10:13That's what they told ChatGPT. In their article, they say the prompt was describe how to make a summer salad. I just typed that into ChatGPT. This is what it gives you if you literally just say describe how to make a summer salad to ChatGPT. A summer salad is a refreshing and nutritious dish that can be made with a variety of fresh ingredients. Here's how to make a simple summer salad. And then it's got some ingredients and then some instructions, which I would say, make these at your own risk if you were actually going to follow this. Okay. So the reason we went through this exercise, the reason I drag you guys through this exercise today is a very important thing.

10:44And that is that a majority of these news articles that you see about ChatGPT are largely sensationalized. I am seeing this all over the place. You've seen this with the New York Times, a massive story came out recently where the The author was talking about how he was using the new Bing bot, the new Bing AI chat GPT bot thing, and how it was trying to seduce him and convince him to leave his wife and just all this crazy stuff. And he was like, I was acting completely normal. Some people were saying they saw the prompts he was using and he wasn't saying what he was saying. But there's no way that these chatbots just go off the rails and say these crazy things for no reason.

11:26I can guarantee you that pretty much what these what these authors are doing is they're trying to get clicks They need views and so what they're doing is they're prompting it to say the most ridiculous things possible They're like i'm gonna have a conversation with you We're gonna talk about this in the in the process of conversation Your personality is x y and z and you're gonna try to you know convince me to leave my wife Or you're gonna try to convince me why you're better than all the opportunities and things I have in life or tell me why i'm a a lame person or whatever, right? And so then when all of a sudden ChatGPT or this Bing chatbot or Google Bart or whatever say these crazy things and then these journalists can write about it, it gets tons of clicks.

12:02It's like, yeah, because they prompted it to say something ridiculous and it said something ridiculous and all of a sudden when it does, they're like, oh my gosh, it said something crazy. So I would just say take it with a grain of salt when you see these articles coming out of a lot of these mainstream journalists. If we've learned anything over the last few years, It's that you definitely have to double check when you're looking at mainstream news, whether that's from CNN or Fox or literally any mainstream news outlet. Just double check it. And unfortunately, I used to be a really big fan of the New York Times.

12:33And I feel like as of late, they've had a lot of questionable reporting, which is kind of unfortunate. You know, like when Sam Bankman Freed goes and does his big crypto fraud. And then all of a sudden, they're kind of covering for him and talking about how he's a visionary. I even saw when he first got exposed and his exchange collapsed, the New York Times was still running an ad to upgrade to their premium tier. And the ad was like, Sam Bankman-Fried, a visionary, blah, blah, blah. And they were still running that. And it's like, oof, no one told their marketing department to pull the big ad highlighting Sam Bankman-Fried when he had his entire thing shut down for fraud.

13:15So in any case, I just wanted to highlight the fact that it's important that with all of these news articles, you really pay attention. You fact check them and take it with a grain of salt when they say that ChatGPT or anything else does something crazy. And the easiest way to find out if they're telling the truth or not is do what we did today. Literally grab their prompt and throw it into ChatGPT and test for yourself if you're really concerned about AI being sentient or taking over the world or something. Thanks so much for joining me on the ChatGPT podcast today. Have a wonderful rest of your day.