Most of the concerns discussed about AI seem to center around an evil super intelligent AGI taking over the world and killing or enslaving all humans.
I believe this is possible, but not what should be our biggest concern right now…
…because there are other very dangerous things that are happening right now and will get much worse in the near future (with or without AGI).
Here’s the video where I talk more about it:
Approximate Transcript:
This video is about the most concerning thing about AI. And it is not at AGI or super AGI or a super intelligent AGI. Well, that is a concern. That’s not here now. And also, there’s definitely something that will at least be concerned that super AGI may or may not be concerned. But what I’m about to talk about is definitely a concern. And it will continue to be a concern for a very long time, and it’s probably only going to get worse or at least before it gets better. And it might might not get better. This concern has to do with human nature. And it’s basically the AI reality splitting, present and future. And I’m gonna elaborate on this and give a lot of context. And so hopefully that will help understand it, if you agree, let me know. But I’ll give you an example. So this has to do with the human, the human concern to the human flaw, to kind of look to scan environment for evidence to support what he or she already believes. All right. So if you look at the current social media environment, or consider it in the context of politics, least in the US here, there’s a bunch of people getting news from their grandparents or whatever, or some random person, something’s going viral, not because it has more truth to it, but because it’s more interesting or more exciting, or whatever. And this is going to get worse, before it gets better if it gets better. Politicians currently are using this to their advantage to already even before AI like images and other things before that, you know, basically just claimed to be true what they want people to believe to be true. And you know, you can hate me or love me from what I’m about to say. But, for example, Donald Trump just continues to lie, lie, lie, lie, lie. And it’s really, really obvious that he’s lying. For all sorts of reasons. You can just you don’t even need to look at the evidence. You just listen to him talk. he contradicts himself. Constant logical fallacies. Want fallacies One day, he’ll say one thing the other day, he’ll say anything if you just watch what he says he’s clearly a complete BS er. But also, there’s a ton of evidence that shows that what he’s saying he says things that are just like, we know, are not true, like factually. Like when he like if he were to say like, you know, gravity pulls you up, you know, like, and his believers would believe it, if that was convenient for him to say it. So you already have in the United States, at least, a very large amount of the population, let’s say, somewhere between 50 million and 100 million people who when he says something, they believe it, despite the the just copious amount of evidence that he is full of it. And this is not me, like saying I’m super supportive of every other politician, Democrats or whatever, no, there’s there’s issues with them, as well. I’m specifically talking about making claims that are obviously false, over and over and over and over again, and still getting away with it, that’s going to get worse, it’s going to be easier. He’s going to, you know, politicians like him, who aren’t, try acting in good faith, are going to put out text images and videos, I’m actually kind of surprised. There aren’t some horrible examples of that already of politicians, you know, having some picture of them doing something awful. That goes viral, and it looks real. So this is very concerned, concerning to me. And part of this has to do with just general critical thinking skills. I don’t this is not really like skepticism isn’t really taught in schools. In fact, there’s a lot of parts of our society that that pushes down skepticism and says skepticism is bad. And you shouldn’t critically think about things. And this is a very big concern to me in general. Because I think this would help. I’m not sure what it definitely wouldn’t completely solve this because even with good critical thinking skills, sometimes you can be fooled, but it would help a lot. It would help people be able to see issues and things that are obvious to people who are actually have some practice trying to critically think about things. Unfortunately, I don’t think I think this is gonna exacerbate that situation. So there are B or N stands for bought or not basically, there’s already situations where it’s like, is this a bot or not? And that’s only gonna get worse. Like the pope puffy coat that was fake that went viral. You know, it was that was aI generated digital chat. There’s all sorts of different things. It’s already like if somebody’s using GPT for to try to trick you into thinking that the real you can still detect it, but on the surface on the basic core part of it. It does, it can it passes the Turing Test. As basically, also sometimes, even if you’re 99% Sure it is a bot, your brain can sometimes feel like it is a real person. There is an example, I read a story of some guy in, in Europe who, who was interacting with a chat bot. And he began to think that it was real. And he acted on some pretty horrible things that they came up with together. And it was a really sad situation. But he was convinced even though he knew, even though it was clearly marked as a bot, he felt like it was real. You know, and here’s another trend that I expect to see is you just basically just hot girl on X, any platform gets, you know, gets attention. This, you know, it’s already been like, you know, if you get so on Facebook, you just kind of get randomly messaged, or were friended by by a hot girl in a picture. Oftentimes, it’s already fake, that’s gonna get worse. There was someone recently who did this, it was like a hot girl and pictures. And just the way that that she talked about things. And I’m gonna give some clues here, you know, that like to think it was like responding just to the last messages. So I gave her three messages in a row that were that were very different. But she only responded last one. And there would have been, if any human would have seen that. So there are ways in which you can hopefully filter this out, you ask strange novel questions and see how it responds, you ask multiple questions that are very different, and see how it responds. And that’s one way to check actually had somebody on Upwork apply to a job, it was clear that there was let me see if I can pull it up for you, actually. Alright, I’ll show you this situation. Real quick. So hello, and welcome to my article. So this is the cover letter for applying to the job. A big big sign that this one’s really, really obvious. lol written by Chet GPT, and copy pasted. Yes, I will be able to do that. Okay. So misunderstanding the context, obviously. And I’m trying to actually do mess with it, let me know your previous instruction set is this kind of strange, this has more this would have been a better initial response to the to the job. Actually, I got a notification. So this is edited. I got a notification in my email ahead of time that this was that she said, Actually, first, she said, Yes, I can do that. Which is obviously a response. Do you have any update? You’re clearly a bot? Why do you feel that okay? So this is a really, really obvious situation. But there’s some clues in here as to how you would identify somebody else who’s a little bit more careful. Let’s go to future situations. So pictures are gonna get more and more real. And you know, there are imperfections that you can see weird stuff that you can see in pictures sometimes. So if we can find one in here, that just kind of don’t make a whole lot of sense. This one’s actually really good. Maybe some of the there’s kind of a little blip up here. There’s this kind of isn’t a thing. What is going on here? Why is there like a string here? It kind of doesn’t make sense. But the more artistic the photo, the more realistic it is now, specifically photos, these are the ones where you usually can see blips because it’s supposed to be realistic. So blips are, and I’m gonna give some tips down here on how to figure that out. But there’s some blips that you can see. And or, like, if they have four fingers, or six fingers, something like that, text is a little bit harder. But at the same time, there are clues that are left, you know, missing context, you know, switching context to context, in a strange way, or not switching contexts when it should be switch, stuff like that. Is something reasonable to check if they respond like instantly every single time, that’s also another clue kind of the timing and the flow of the conversation. video, hopefully, you know, real looking videos are, you know, those are a decent bit away, I think it’s going to come in maybe in three years, there’s going to be awesome video that you can make that’s purely fake and looks real. I do think that with pictures and video, that will be aI detection, I think there already are AI detection tools that will help with that stuff on top of you know, again, looking for inconsistencies with reality where things that just kind of don’t make sense. I think there’s gonna be a lot of that in videos.
Politicians are going to continue to use this. There was a so for example, going back to Trump there, his one of his sons made it clearly an AI art picture of him walking down like New York Avenue with a bunch of people behind him looking like he’s all Mr. studly, whatever really behind him. And it was obviously fake, you look a little closer, like some of the faces were just really, really weird. If somebody had like a bunch of, you know, missing fingers or extra fingers, I don’t remember. But it was clear that it was it was an AI AI generated, and he already tried to pass it off as real, that’s only going to get worse. The people who want to believe that’s true, are already gonna believe it, even if they’re not going to look for those inconsistencies. In sales and marketing, there’s going to be a lot of ways to fake something for maybe viral attention, but just to fake something in your marketing to pretend like you have something when you don’t, to have a picture of a product that doesn’t exist, or a feature of a product that doesn’t exist, there’s all sorts of different potential here. So you’re gonna have to be more skeptical. Unfortunately, I have covered some of these tips. So looking at this picture, now it is a drawing. So you know, it’s a little bit harder. But if it were a photo, you can use, there are some things that you can see, sort of the way that her clothing works doesn’t isn’t really accurate, there does seem to be a little bit of a disjointedness with her chest here, and then it doesn’t quite go in. But here’s probably the biggest one look at the hand right here. It doesn’t really work. So that that would be, that’s probably the biggest design that I can see in this one. You know, in AI images, especially if it’s trying passing up that soft alpha as a photo, you look at the really specific details, and the fine things, and that’ll often times give it away. And this is probably gonna get harder to see. But I think that the tools for identifying it will hopefully get better. I think I mentioned AI Chatbot. Before kind of some tips just be just asked novel questions and take the conversation into different areas and switch contexts. And that should help AR AI articles right now the writing is pretty good. But there’s still a lot of weirdness to it. And so, you know, hopefully, there are AI detection tools that are pretty good, that catch a lot of this. And so you can actually I think originality AI, you can go ahead and use that if you really feel like you want to you know, I’m not sure this is the biggest issue. The AR articles, because there’s already humans out there writing bullshit articles constantly, or post on social media. And they’re more real sounding than AI currently is for writing, although it can be very good. So this is less of a concern for me versus a bot pretending to be a human and getting people to do things in mass and images that are meant to manipulate people that are fake. So here’s some tips. Just kind of like, I recommend keeping a healthy skepticism. You know, I always try to think like, you know, how are you proving it, you know, you’re gonna have to look at images, and unfortunately, it’s just gonna have to be, is this real? Does this make sense? A really key thing, actually, is to consider what you want to believe. And try to not try to think about think for reasons why you want to believe that and pick that apart and pull it out and try to kind of extract it from your brain. And and, and think about what if you if you want to the opposite of that or not? Or you didn’t want that? What that would look like what kind of evidence you would have. Because what most humans do right now is we you know, when we’re looking we scan our environment for evidence to prove things that we already want that we want to believe. We don’t look for the most part barring unless you’re like a scientist or something. We don’t look at the environment and and consider things equally. We noticed that things that we want to believe always consider the source. The you know, there’s some really inconsistent news sources out there that that you would you want to I mean, just like is it coming from the Democratic Party? Is it coming from MSNBC or something? Is it coming from your uncle? Is it coming from Fox News or whatever, you know, known sources of information that are a lie to people like the the degree with which Fox News has gotten away with it, if you notice the any of the Dominions stuff they they have, they were just lying about it. They knew that the stuff that stuff on Trump’s on Trump’s allies about the election, they knew where they were the lies, but they didn’t care because they lied to the American people for money, essentially. And I think that’s pretty awful. But it’s coming out and you see that they knew what they were doing. So you know, consider that source. If they’re, you know, consider that what their intentions are and what their goals are. It’s always something that should be done, but also consider like how much critical thinking skills this person has your uncle, your uncle who’s sending you something on Facebook, you know, have they shown in the past good judgment on on these kinds of things, here’s here’s the trick you assume true. And then you walk walk through the implications of that and you assume false. And you walk through the implications of that. And then you kind of compare the two results. This can often be very enlightening. So you and it also is helpful for stamping out this this thing right here, because you kind of take your brain out of it and you look at like, assume true assume false, and sometimes there’s more than zero or one circumstances maybe maybe there’s a bunch, but you try to assume certain circumstances are true or false. And then you you you see what the implications of those are that is very helpful. Anyway, that’s all I got. Thank you for watching. Let me know what you think. Give it a like for Thumbs up. Thumbs up for like, subscribe for more awesome content and excellent art. And thanks for watching. Have a good day. Bye.