the business, the owners off next is looking at the rise of a i and if it's to regulate, isis is more on the t w dot com to jar draining, building. thank you for watching the can use. yeah, we are all set. we are watching to see all the to bring you the story behind the new we own about on volume information for free might do to name done name is the calls back. said wow, thank you so much for joining in. welcome to don't hold bad. a lot of people do that. it's all about saying it aloud. you guys, would it be nosy bay?

like good everyone to king, to check out the award winning called. com. so hold back. the artificial intelligence is transforming our world. and the race for domination has the gun from shaping the future of search engines to driving. the next era of warfare a i, it's fueling, and new industrial revolution is becoming less than the fee for everybody. so for something in this video, we'll look at where the power lines will as test the role of government and the task. right now, it's a big divide between the haves and have nots, the stakes of falling behind. so for military's, if they get this wrong, they use a war and the challenges of regulating an industry that is growing ad, mind boggling speed. i think

a lot of people who are in position to make policy don't really understand the technology. that's all coming up on business beyond this let's get this side of the way 1st. there's no universal definition of artificial intelligence, but some experts say it's best described as a system that can take action on its own. self driving cars, high frequency trading on the personalized content that appears on your social media feeds are all examples of a i as a recency launch language models like microsoft box top g p t. google's barge, they do is ernie and mess of lama. we'll talk more about those in a moment, but 1st, let's follow the money. since 2018 revenue from a software has increased rapidly, with the market, expect it to be worth a $126000000000.00 by 2025. venture capital

investment in general is of a i the kind of systems that power those top box we just mentioned have seen especially large growth over the past few months. generally of a, i has been robbing the headlines when open a i released child c, p t, millions of people including myself or wired bias requesting abilities. but how big of a deal is it really? i asked a researcher at melanie mitchell. i think it is a game changer. you know, it's really enabled uh, a i systems to generate language to generate images and soon to generate videos in a way that it's never been able to do before. that really can be quite human, like the world of big tech took a similar view. all of a sudden, people were wondering if a child g p t powered version of microsoft being could be about to topple google's domini, ends in search. but this is about a whole lot more than

a few tech companies trying to do each other. the way generative a i is integrated into us search engines will have a profound influence on how we access information. and by extension on our understanding of the world, is this big implications for who gets to decide? what is the future of a i, one of the things that we've seen with large language models i've tried to be, is there a lot of problems they can generate text, that's really impressive. in many ways it can be very convincing. it's very good quality text. but because it's trained on the data on the internet, the database of these models has all of the problems that exist in the text on the internet. so if prompted in the wrong way, these models can produce text, it is hostile toxic or bias to difficult ways. the models themselves tend to make up facts or hallucinate things. in some cases,

things can get done right. weird in this exchange between things a i model in, in new york times reporter the truck bought, appears to pull in love and tried to convince the reporter that he doesn't love his spice. they sounds like humans. you know, you ask them questions and they answer in a very articulate a fluent way. sounds like you're talking to a human in some ways. and yet they don't have the understanding of the world that humans have. that enables them to be robust and to tell the truth, to be able to tell the truth from a non truth from misinformation. and that's something that humans can do much better than machines. at this point. in recent years, the field of ai has moved away from being primarily an academic discipline. one big reason for that is funding. so developing a language model like taught

t p t takes years and relies on massive look tricity goals, link computing systems. the reason why this very computationally attention approach has effect on global power is that it's very expensive, trick costly to assemble this kind of nasa of computing centers that are then used to train these machine learning models. and that's why it's large tech companies like google and microsoft that we're seeing releasing these early commercialized versions, for example, of large language models incorporated into their search functions like google search or bang rather than just any research or able to do it. so it seemed that big tech, with its deep pockets, holds a fair amount of power in the world of a i. but how does that translate into national pro wes, when it comes to regulating a i am the digital world more generally. the european union is at the forefront,

the e u. a i x expected to go into force in 2024 attempts to categorize a applications according to their risk stations. applications such as the social scoring system, the chinese government has introduced would be deemed an unacceptable risk and therefore banned high risk applications such as c, b scanning tools or shot bonds would be subject to certain legal requirements. busy other applications like gaming would be left largely on the regulations. the us in europe, a very different philosophical approach is when it comes to tech regulation. we certainly see with data regulation like the g p, r, as well as all the areas like a more broadly e regulation on artificial intelligence that e is leading into regulating technology. whereas the united states has much more of a less i fair approach. while the you may be ahead in the area of galatians,

there are a number of open questions, especially when it comes to how the law will be enforced. researches also point side, but there is no way of knowing how public institutions are already using ai is layla, such as describing the situation in germany. so for example, that mom time a german city is you, they face recognition software in order to help the police with being notified and, and certain kind of risk. and that situation, we know that this application isn't working with a but rather with, with very fast that body movements. so this could be an example of that showcase how base trig transmission security issues, but also data protection issue issues come together. another ethical dimension to the story and one which has to do with power to concerns the people employed, to train the machine systems open a. i does not give information about how many people were used to train chop g p t

. so i us a researcher, melanie metal for her best gas. yeah, that's hard to say, you know, hundreds to thousands of. there's human involvement at all different levels. there's the human involvement in of figuring out what data to give the system to learn and you know, what, what texts and so on. there's human involvement in training, it to be kind of more conversational and to answer questions in a more human like way. and then there's a humans who try to prevent it from spewing out toxic language or other kinds of things that you know, the companies might not want it to. to say, since the release of chat t p t, an investigation by time magazine. find that open a i used workers in kenya to classify disturbing and legal content, including descriptions of sexual violence. employees reported receiving in adequate mental health support and earning less than $2.00 an hour. i think it's pretty

obviously a monthly, problematic to outsource a to, to countries. i mean, i have no problem with as are some of the countries that, that i'm for making everything norman's finest of money on the, on the exchange rate. luckily, i only sense that she needs the cheapest to buy to a treat. she actually go home and make sure that they have adequate health care, which again is presumably pretty low cost for the car. in this video, we've looked at the rapid growth of the industry and how the focus of power can shift between big tech companies, governments and research institutions. but not everyone i spoke to agreed that framing the current status as a competition make sense. and every time i call part regarding a i is definitely farming to the discussion when the quantity of discussion around and course make decision making. so when i hear, let's say that your needs to catch up chinese or american

division and position regarding the use or decline development. okay. i see that we haven't really defined what the goal of this phrase would be. the other is believe that distinguishing between democratic and, or for a terry and regime is essential in any new digital order. i'd like to see the developments something like a, a shrink and area among a life. scientists, making it easier for a scientist to travel between, say, europe, united states, australia, japan, south korea, and make it easier for them to move around, share information. but one thing that blew my mind was when a i research or a melanie mitchell told me that no one actually understands how a i systems like touch e p t actually work. or why they come up with the on to is they do? i asked her to clarify if she really meant no one, not even the scientist to program that. that's right. uh, you know,

the systems are these very um, large computer programs that stimulate sort of what they're called neural networks . didn't her own it works, and they have stimulated neurons that have connections that have a weights values. and they have something on the order of, you know, billions to trillions of these weights. and that's just an unimaginable scale for humans to understand. now they were kind of like the way we might try and understand our own brains, which we don't. the spread of misinformation is frightening enough, but in other sectors, the trends of unrestrained a i could even be access dental. i think probably the people are the most successful in the arab solution, although as you find ways to use these technologies to be more productive for me, more competitive in the workplace. but for military's, if they get this farm, they shall use a war, uh, their country. we could, you know,

lose power on the stage that could be overtaken by the arrival they could be invaded. and that brings us to the end of this edition of business the. we'd love to hear your thoughts on this topic. so to leave the comments. and if you like what you see here, you can always subscribe to the business beyond playlist. a good place to start would be our recent video and looking at whether the dollar will maintain its position of global dominance for me on the team. it's good bye. take care the the 12 candidates 30 meters above ground. challenges simply and push

them to their limits. brewing company is breaking new ground on the job market. attracting applicants with a trainee contest who will make it. and guess the apprenticeship made in germany next on d, w. facing internal and external friends. and nation in the shuttle, the things are taxed on ukraine has deeply divided the population some once and alliance with others are returned to rushing in 45 minutes on d w. what secrets lie behind these discovered new adventures in 360 degrees and explore fascinating boats. heritage

dw world heritage 360 now the did you know that dancing in a club can have real monetary value not just from admission sales and expensive drinks. a club has massive energy costs to carry on a weekend. it consumes as much energy as a single household in a year. but it can also when back these costs with every step you take, house will show you in a moment. the . 3 sun wind. 3