The US Government Just Changed EVERYTHING (New AI EXECUTIVE ORDER!)

So this is probably one of the most important videos that we will make this year. And this is because this video is regarding the executive order that Joe Biden just signed. And the reason this is one of the most important videos of the year is because this executive order will likely dictate how the laws and legislations are put in place on AI safety and regulation. And that is about to change the entire landscape that you think is coming. I mean, there is so much to discuss because I've seen tons of people talk about this, but they haven't got into all of the nitty gritty points that we do need to discuss. And I'm going to be paying very close attention even more to this situation as it develops, because the way how this situation transpires could change the entire future of artificial intelligence. So let's get into it. So if you haven't heard already, Joe Biden tweeted out this one thing is clear. To realize the promise of artificial intelligence and to avoid the risk, we need to govern this technology. There's no other way around it. Yesterday I signed an executive order that is the most significant action any government has ever taken on AI safety, security and trust. So what is this executive order? Basically, it's how they're going to take take care of the AI situation. AI is a new and unpredictable technology and they need frameworks in order to protect the general public. And of course, they need to regulate this kind of software, this kind of new technology. And the problem is, is that it's developing so rapidly that it's developing much faster than the laws can adapt to it. So what they released as well is they released this fact sheet. And the fact sheet is essential, essentially a type of information sheet which contains the Cliff Notes version of the full PDF. So I'm going to go ahead to that page now. You can see what we have here is executive order on the safe, secure and trustworthy development and use of artificial intelligence. So this right here is a very, very long document. As I'm scrolling down, you can see that it's literally huge, huge, huge. But if you don't want to go through this entire document, you could of course ask Chat GPT to summarize it, or you could just go through the fact sheet. And there are certain points that we do need to discuss because the stuff that's happening is absolutely incredible. And there are some more discussions going on on Twitter right now that show us that many people who are leading these companies feel a very different type of way. So what are the main points from this executive order, and what do we need to understand and as to how it's going to change the landscape right now? So one of the most important things that they do talk about is political declaration of the responsible use of artificial intelligence in the military and autonomy, so it says. In February, the United States made a political declaration on the responsible use of AI and autonomy. And they are announcing that 30 nations have joined the United States in endorsing this declaration of calling on others to join. This declaration establishes a set of norms for responsible development, deployment, and use of military AI capabilities that can help responsible states around the globe harness the benefits of AI capabilities, including those enabling autonomous functions and systems for their military and defense establishments in a responsible and lawful manner. The Now I decided to highlight this part of the fact sheet because I think it's one of the most important parts. We all know that artificial intelligence is advancing at a crazy speed. And with that, the military applications are going to be truly incredible. We all know that one of the reasons that the military is particularly harrowing is because humans have to make decisions on how they're going to kill other humans. But with robot pilot fighter jets or whatever, whatever autonomous drones swarms that are going to be deployed in this next era of artificial intelligence warfare, what are going to be the guardrails? Are there going to be certain laws and certain norms and certain systems that are going to be deployed, or are there going to be certain regulations that prevent them from doing this? Now, the reason I decided to also talk about this was I'm not sure if this entirely makes sense. Now, I do know that it in war, there are certain things that you can't cross. There are certain laws and certain guidelines. But how does this make sense if you are against war with another nation? Now, I'm not sure how any future wars are going to transpire, but let's say, for example, theoretically speaking, we know that the United States, you know, not theoretically speaking, but we do know they spend a whopping $877 billion annually on their Department of Defense. Now, what's crazy about that is that we do know that the United States military works on top secret things all the time. So how are they going to say we're going to establish norms for the use of military AI capabilities around the globe? Because I think that the United States is one of those countries that knows AI development is something where you're either in front or you're vastly behind. And I don't think the United States is ever going to get behind on this race or even wants to be a moment behind. Because if you are a moment behind on AI defense, I'm pretty sure that you could essentially lose this war. I do think it is good that we do need to have this where we do need to talk about the set of norms for responsible deployment and development. But I do think it is going to be a bit difficult to do so when many nations that are working on advanced technologies are always doing so in secret. So I'm not sure if this is going to work. I'm not sure how they plan on doing this very effectively. Because whichever country is on a top secret program that has a clearly advanced AI military strategy, they aren't going to be sharing it with other neighboring nations, especially those that they may have to come into conflict with later. So this is why it's really interesting because if you know that United States, you know, whether you think Area 51 is real, whether you think it's fake, every country is working on top secret military stuff. And I think it will be interesting to see how this development does occur. Now there is some other developments. It says that the US Air Force is moving fast on AI piloted fighter jets. After successful autonomous flight tests in December, the military is ramping up its plans on bringing AI to the skies. So you can see on the morning of December 1, 2022, a modified F16 fighter jet codenamed Vista X62A took off from Edwards Air Force Base, roughly 60 miles north of LA. Over this short test flight, Vista engaged in advanced fighter maneuver drills, including simulated aerial dogfights, which before landing successfully back at base. So while this might seem like business as usual, it was not a fighter pilot at the controls, but for the first time on a tactical aircraft, a sophisticated AI that is absolutely insane. And we did talk about this in a previous video. We talked about how AI fighter pilots are likely going to be the future of aircraft because they are going to be able to respond instantly. They don't have to be subject to the G forces that pilots have to do. They don't need additional training. You also don't have a human pilot that is going to be subject to any kind of death or physical industry. And potentially when engineering these pilots, you're not going to have to engineer a seat and of course a parachute that is going to be ejected on the cockpit. So with these things, I think this is definitely the future. Because additionally you could also train an artificial intelligence in certain simulations like Nvidia's got state of the art systems. They could develop one for the military where they Train planes in one. And it could see a billion or a million different scenarios with this and it's likely never going to use. Imagine going up against a computer that's seen a billion different scenarios and it's trying to evade you is going to win 100 out of 100 times. I truly do think that this is what we're going to have because at that point, human pilots are going to be severely disadvantaged. Then of course, we have detecting and blocking AI driven fraudulent phone calls. The Biden Harris administration will launch an effort to counter fraudsters who are using AI generated voice models to target and steal from the most vulnerable in our communities. The White House will host a virtual hackathon and inviting companies to submit teams of technology experts focused on building AI technologies to come together and build AI models that can detect and block unwanted robocalls and robotechs, particularly using those novel AI generated voice models which particularly harm the elderly. Now, I do think that this is going to be one of the biggest problems in the future. As we do know, the AI voices are absolutely incredible in their current state. The only problem with them which is currently being solved is the fact that they don't respond almost instantly. And this is something crazy because the clip I'm about to show you was one from five years ago. So Google essentially did release something, well, didn't release something. They actually spoke about something at a, you know, conference or hearing or whatever. And they talked about, okay, an AI assistant that could now make fun of phone calls. And this is why everyone in the AI community says that Google has dropped the ball. So I've titled this probably why Google hasn't released this and never will. It's because Google's AI assistant is absolutely insane. So I'm going to go ahead and include a clip of that and then when you see that and you hear that, you'll understand that if that was five years ago, Google is sitting on that kind of technology five years ago. Imagine the kind of phone call technology that they could have now.

So let's go back to this example. Let's say you want to ask Google to make you a haircut appointment on Tuesday between 10 and noon. What happens is the Google assistant makes the call seamlessly in the background for you. So what you're going to hear is the Google assistant actually calling a real salon to schedule the appointment for you. Let's listen.

Oh, how can I help you? Hi, I'm calling to book a woman's haircut for our client. I'm looking for something on May 3rd sure, give me one second.

Mm hm.

Sure. What time are you looking for? Around at 12pm we do not have a 12pm available. The closest we have to that is a 1:15. Do you have anything between 10am and 12pm depending on what service she would like. What service is she looking for? Just a woman's haircut for now. Okay, we have a 10 o'.

Clock.

10Am is fine. Okay. What's her first name? The first name is Lisa. Okay, perfect. So I will see Lisa at 10:00 clock on May 3rd. Okay, great. Thanks. Great. Have a great day. Bye.

So now that you've seen that clip. Okay. And I'm pretty sure if you've been a subscriber on the channel, you probably would have seen that clip sometime before now. With that being said, isn't it crazy that in the future these phone calls are likely to get better? And imagine now a hacker or someone who is a friend fraudster having access to this kind of technology. Even if Google open source this technology or even if they didn't, I'm pretty sure hackers and fraudsters would find a way to use this software to jailbreak it, to get it to say what they want it to say. And of course that would pose a huge, huge problem. Imagine everyone's phone, just automated phone calls, just that sound like real people and interact with you on a real, real level. That would be absolutely incredible. And I really hope that is in the future because sometimes you really do know or sometimes you don't know who is calling you. And of course, especially the elderly, they have no idea. And of course we have something from Geoffrey Hinton. And it says, in my opinion, the public perception that Google is not making much progress in AI is wrong. They invented the Transformers and diffusion models. The perception comes from the fact that they have been very cautious in what they release. And as you can see, that is very true. I and this was Geoffrey Hinton responding to someone talking about the fact that essentially everyone is believing that AI is far more dangerous than it is. And then he was responding in another comment saying that, listen, I literally left Google to talk about how dangerous AI is. So I don't think it's one giant conspiracy. But as you know, this is exactly what we're talking about. Google have been so cautious to talk about what they've released and we've seen by the, you know, the previous clip that it's true. You know, I wonder what else Google is sitting on. And that also does bring us to another question. Okay, so now that companies are going to be bound by certain, I guess you could say, laws and legislations to where they need to share what they're working on. Will companies just pretend that they're not? I mean, if I was the CEO of a huge company or if someone else was, I can't imagine that they would happily give over some exclusive information straight away to the government. I mean, why would they? I mean, it's kind of interesting to see how this is going to unfold. Then of course, we have something really interesting right here. We have international norms on content authentication. And I think this is something that needs to be solved rather quickly. The Biden Harris administration is calling on nations to support the development and implementation of internal standards to enable the public to effectively identify and trace authentic government produced digital content and AI generated or manipulated content, including through digital signatures, watermarking and other labeling techniques. This effort aims to increase global resilience against deceptive or harmful synthetic AI generated or manipulated media. And essentially this is really true. And they're doing stuff by, I think, collaborating with 15 leading AI companies to develop mechanisms that enable users to understand if audio or visual content is AI generated. And we've all seen AI generated stuff floating around the Internet. We've heard voice notes from your favorite celebrities. We've seen figures like the Pope in certain jackets, which of course aren't real. This is AI generated, although it does look extremely real. But this is a giant problem. Okay, earlier this year, did you know that there was a picture that literally managed to crash the stock market? A few hundred points. Maybe not a few hundred points, but enough points that billions of dollars got wiped out. I mean, imagine what's going to happen when we have videos in the future, when we have phone calls in the future. We could have mass media at an insane level of deception, which could present a whole host of problems. But it isn't all doom and gloom because some companies are working on this stuff. But there are a few problems that we are going to talk about regarding this. And it was interesting doing the research for this because I didn't really know that these companies were working on this groundbreaking technology. So take a look at this. Okay, so it says Google DeepMind has launched a new watermarking tool that labels whether images have been generated with AI. The tool, called Synth id, will initially be available only to users of Google's AI image generator, Imagen, which is hosted on Google's cloud machine learning platform, Vertex. So users will be able to generate images using Imagen and then choose whether to add a watermark or not. The hope is that it could help people tell when AI generated content is being passed off as real or help protect copyright. Now this is something that is really cool. You can see that. This was published on 29 August 2023. Identifying AI generated images with Synth ID. It says new tools help watermark and identify synthetic images created by images. So essentially they show us this video where they pretty much say this image tool is new available here, it's available through this. So with Synth ID you can add a watermark to your image and it can be detectable even if edited by common techniques where people, you know, try and crop an image or where they try add advanced filters and it's something that is really, really cool. However, although this is a good system, it doesn't work that well, apparently. So this says that researchers tested AI watermarks and broke all of them. A research team found it's easy to evade current methods of watermarking and even add fake watermarks to real image. This is an article on Wired and essentially they talk about this and it's pretty scary because he said there's no hope. Okay. And I don't want to be an AI doomer, but I do think that eventually we will be able to solve this image generation thing. But the problem is, is that I do think that, you know, the bad actors who do want to, you know, create that level of distortion or create that level of media distortion, they're not going to be using models that, you know, have AI generated watermarkers on them. They're simply not going to be doing that. And I do think that it's going to be increasingly hard to solve this problem. So whoever does manage to solve this problem is definitely going to be someone who is really, really wealthy. So for then of you who are out there who are going to be starting an AI company, this is a problem that does need to be solved. Then of course we have pledged to incorporate responsible and rights respecting practices in government development, procurement and use of AI. So basically this is just talking about the fact that when governments use AI and when they deploy AI, they need to do it in a responsible manner. And this is really true because if they don't, this could be used in a billion different ways to invoke a total, to invoke a totalitarian state which essentially means dictatorship, like control and stuff that you really, really don't want. So I hope they do focus on this. But then again it's the government and do governments really govern themselves? I mean, I don't know. It's something that's up for debate. Politics is really tricky, but it will be interesting to see if they do this effectively because if they don't, then certain countries could definitely get out of control with their generative AI use. And there's just a million different ways that you can use this. In terms of all the scanners that are now coming out, all the advanced stuff like, trust me, guys, this is something that needs to happen. Now. This was something that was pretty controversial. It said President Biden has signed an executive order that require AI companies to address algorithmic discrimination. Now, I think everyone's missing the point. Some people were saying that, okay, certain groups are going to get preferential treatment, certain groups are going to get less treatment. It's saying that why do we care about this, yada, yada yada. And I think people are missing the mark about this, okay? Because although the social lens does flip from time to time, what we have to understand is that this isn't the main point, okay? What you have to understand is that the problem with AI, well, not really. The problem is that. And this was mentioned in a video that I was watching the other day on a very good channel and essentially what this guy was talking about, and this guy is David Shapiro. And one of the most important things he spoke about in this video was that our chatbots lying, okay? So regardless whether you're left leaning, right leaning, conservative or libertarian, the problem is, is that now we live in a different state. Okay? Previously we with the Internet, what he spoke about was that with the radio and stuff like that, the Internet changed everything because you can go on the Internet, you can search for information and then you can decide for yourself. So the Internet removed all barriers of communication. This is true, okay? But now the problem is generative AI and ChatGPT. You can have deliberate deception. We can have people at OpenAI or people at Claude or people at Google decide what is right. These AI systems, they can essentially decide what you are going to be told. Okay? So he talks a lot, a lot, a lot about perverse incentives. There's just honestly so much. This video is really detailed and it goes into it, but it brings us to the point that whatever these models are trained on, whichever inherent bias that they do have, you have to understand that these models are going to be worldwide, they're going to be ingrained in many different systems. You have to think about how many different systems are calling on the GPT Force API to. That basically means that whichever large language model is going to be the front runner is essentially going to be the oracle of truth for what many systems do use. And that's a bit scary because if they don't disclose what that is, then it means that they could dictate how certain things are perceived and how certain issues are essentially interpreted by individuals. And I think something like that is definitely, you know, more interesting because rather than, you know, most people, I think in the future, rather than them searching, okay, is the sky blue? They're probably just going to ask the app on their phone, is the sky blue? And it's going to say yes. Now, of course, the sky being blue isn't something that is controversial, but with other controversial issues, I think it's going to be a bit more confusing. And in this video, this is where he talks about deliberate deception by air companies, where they're going to, you know, withholding information, deceiving or misleading users for legality, morality. There could even be financial gains for them to, you know, mislead people. You know, maybe what if one of these companies wanted to mislead people on certain information so that they could financially profit from it? Or what if, let's say, for example, a billion dollar company, previously, if you know about society, they used to say that cigarettes were good for you. Okay? And you know, you could have these cigarette companies pumping money into these generative AI companies saying that, hey, make sure that your chatbot says that, you know, cigarette smoking is good for you. That's just an example of something that could happen. But of course, you know, not right now, but of companies that are out there that could definitely misconstrue that data for financial gain. And it's going to be something that is definitely pretty scary. So we have to hope that whatever training data that they use and whatever weights and biases that they do have, they're not overly, you know, leaning to one sort of corner because this could present just a whole host of different issues. So, yeah, it's definitely good that there's a different loads of models to choose from, but it's not great that there's only one that everyone uses because with deliberate deception, it could result in a whole host of other issues. Then of course, we have one of the other points that many people didn't think I was gonna bring up, but this is a really good point. Okay? It says, in the name of safety, a US executive order passed the limit to AI to require the most powerful models to register with the government users and kyc, the most expensive AI laws in the world. The countries that don't do this will be the winners of the next century, there will be no catching up. And essentially what this is, and I think this is also really important, is that they talk about, and many different Twitter users were talking about how generative AI and what they are doing now is they're going to be tracking companies that have a certain amount of GPUs and they're going to say, look, if you buy a certain amount of GPUs, you need to report that to the government. Now, the number that they did use, currently there's only two companies I think, in the world that do have that amount of gpu. So it isn't aimed at everyone or small open source and foundation models that people think it is aimed at. But people are stating right now, and I think this is the wrong argument, people are saying that the US Is going to stifle innovation. I think it's going to be the absolute opposite. I don't think the US Is going to stifle innovation at all. I think the US just wants to keep a close eye on what's going on and I think they want to stay ahead of everyone else because we have to understand that United spends billions of dollars on making sure that the greatest country in the world and most advanced. And I don't think they would ever lose that in the air race, which is why they're making these laws right now. I think they just want to make sure that whatever they do, it isn't going to other nations. Because I don't think the US Is going to say, look, we're going to just regulate everything down to the core and let all these other countries go off. I think they know that this is a race. And I'm pretty sure that when they look at the big picture, they, they are going to be able to see that. Well, at least I hope.