then you can find much more on the website, that's d, w dot com. ah, artificial intelligence is constantly making our lives easier. chat g p t can write us essays within seconds. siri helps us organize our lives and algorithms, hand feed us personalized content on social media for me. that's any kind of food porn. but along with these impressive advancements comes the concerned that these machines are getting too powerful. so where is i taking us? the potential risks and future of artificial intelligence. that's our topic on shift. ah, there are 2 types of artificial intelligence. there's the, i from sy fy movies where machines are able to pass off as human. and there's the i, which we actually use right now in everyday life. and it can do more than sort data analyze patterns and automate processes. for example, chat j, p t,

which is the talk of the tech world at the moment. if you haven't used it yet, give it a whirl. and while it's doing some really incredible stuff, it's also making mistakes. plus, there are concerns that this type of a technology could be abused. chat t p t can do anything from draft e mails. to brainstorm answers to questions on a job application. the a i bought has even passed the u. s. medical licensing exam, g p t for the newest model developed by albany. i can do even more the i ipad is capable of tackling more complex task like doing your taxes. not only can g p t for process text, it can also work with images hitting the ai, a photo of the contents of your fridge, for example, will enable it to suggest a recipe for dinner. you can also sketch a website mockup and g p t for will transform your drying into working code. but there's a flip. i'd thoughts like chat, g, b t make things up at times. it's

a phenomenon called data hallucination. sometimes when a i bots don't know an answer, though, referred to person's figures or studies that don't actually exist. a i chat bots learn a lot from the internet. the consequence is they can be prone to reproducing, hate speech and racist or sexist stereotypes. ernie bad from chinese tech giant by do for instance, was reluctant to answer questions and chinese president chin ping, where the countries covered 19 policy in its 1st public tests in march. there's no doubt about it. a i bats can be manipulated and misused. they can be programmed to avoid certain topics or to spread fake news. and authoritarian regimes could use bats as tools to influence public opinion. as you can see, i can be problematic, and the tech world is developing at a rapid pace. check out these photos. it's the pope sporting

a white puff at jacket. i look pretty real, but they're not. they were created with an ai image generator and posted on twitter examples like this show just how easy it is to produce fake and misleading images. there's been talk about mass producing misinformation online, which is currently mainly done, you know, manually in the sense that there's a large numbers of people in so cultural factories that produce misinformation and intervene in online conversations on social media and elsewhere. i've this, if this can be and when this can be automated on a massive scale, that, of course there's a, there's that kind of an whole new level of concern. not exactly what we want from ai. we'll makers are struggling to keep up with the rapid developments. regulation is a must even some developers, a calling for it. and the european commission is now proposing a legal framework to tackle the risks of ai ah,

the european commission wants to take a risk based approach to ai. the framework categorizes a i systems and proposes regulations accordingly. level one is unacceptable risk all a i systems that are viewed as a threat to security, livelihood or human rights are supposed be banned. that could include anything from government, social credit systems, to toys with voice command that could lead to dangerous behavior level to as high risk. these a i systems would have to adhere to strict regulations, take, for example, scoring systems that could determine a person's access to education or a i used in robotic surgery level 3, limited risk. this level is applied to most shut bots and makes transparency mandatory for the developers. ah, level for minimal risk. a i systems like spam filters, for example,

wouldn't be regulated the art world is already reeling from advancements in artificial intelligence. text to image tools such as dolly and mid journey are generating pictures that can pass as human made photos and paintings. as artists and creatives face competition from our systems. lawsuits are already popping up very good at doing this sort of things. you might get a graphic designer to do and the guy good at doing some fun things. the things that putting together, you know, a koala on top of a boat or things that you might, sir, have never seen before. these types of a i platforms, use images from the internet to train their systems. issues related to copyright and ownership have largely gone and addressed this is a really tricky question. and i think there's already ongoing cases in, in a court, in courts of law trying to sort of set the set the borders of the boundaries.

photo agencies, getty images is suing the company, stability i for using its photos to train its software. generated a i draws on examples from the internet. these systems are trained with work from real artists. some feel the threat to their very livelihood. but not all experts are pessimistic. oh, i don't think he's going to take away the jobs of voters. i don't think it's crated in any way. replace out is just like photography did remove painting what these programs produce may be useful for illustrating articles like but they're not going to speak to us. the way that great out speaks was that's because, unlike great art, i can only create images using existing goddesses. and these systems need a lot of practice to create a good picture. the crucial question here is, is i are working for us, or is it taking work away from us? a i, systems are adept at performing automated tasks,

such as filtering information out of large datasets. no profession is really safe. so the moment we have in our aspect of our work that are about reproducible patterns, this part can be reproduced very easily by machines. the capabilities of a i are expanding. i systems can now produce creative content such as essays and code experts predict that going forward. many more tasks could be taken over by a i to people with actually very little skills can just by using some prompt or inserting some prompts into an interface, produce an art pictures and novels possibly videos. so it, it's also these, these creative jobs that are increasingly being threatened, i would say by automatic reproduction. but new i systems still require human oversight. standard a i tools can streamline our work,

plus new jobs will emerge in this rapidly developing world of a i, for example, data detectors who check and analyze the training data for a i. systems. prompt engineers who work to improve a i tools by fine tuning and optimize and text prompts. educators who can teach the skills needed to work in a i, she learning. and of course programmers will remain in high demand. there are certainly ways in which existing jobs can benefit from ai and there's no doubt about it. these advancements are impressive, but i, i today is still no where near what some see as the ultimate goal machines that can think like humans and learn independently. this is known as artificial general intelligence. a i does anything from filter emails for spam health smartphones understand human voices or screen x rays for cancer. but most systems are only good at specific tasks. unlike humans, who have skills in a wide range of areas,

human, radiologists don't just bought cancer, they also understand language. they can tell if an email's spam. they can also learn how to write a bike, play the violin, read a book, the list goes on. the idea behind artificial general intelligence or e g i, is to create a computer system that can learn and perform the same vast amount of things as humans. icelandic research or kristin thorsten is one of the people studying this type of artificial intelligence. so what does he hope to see in the future? ah, one of the main reason why i chose to dedicate my professional lifed to a i is because i thought that i would be able to see some amazing things in my lifetime. at the top of that list is realizing general machine and totals. and i absolutely don't think that is impossible, but it's going more slowly than i thought. just precisely where opinions diverge,

hears philosopher nic bostrom. it could happen surprisingly soon. menda could be somebody making some breakthrough a few years from now and suddenly fall takes off. but this breakthrough hasn't happened yet, but the researchers are more cautious with their predictions. 40 to 60 years. i'm not came to their predictions about that. they don't just extremely likely to be wrong so far, the extent to which i, i will change our lives is still pretty uncertain. but one thing is already perfectly clear. it will definitely be part of our future. so it's hard time we as a society address how far we want these developments to go and what we should be cautious about. the prospect of systems with human like artificial intelligence, stir our imagination. some are more excited by the potential. others are more wary . some think a i systems could become our most loyal companion and relieve us from the burden of work. but others warned that only

a tiny elite could end up benefiting from the tech. and once the system surpass our intelligence, they could turn on us. we have to ask ourselves 1st the do we want to live in a world where humans share they wold with entities that are not human. but that can reason better than they can and that can reinvent themselves faster than human scott. and then we have to ask, who will decide it should be a global conversation, because this is about humanity as a whole. this isn't only about societies relationship to ai. these developments can also have implications for how we, as humans interact with one another. what i'm concerned about is that some humans will be considered less human than this artificial intelligence less deserving of rights, less deserving of proper treatment. because that is a trend that you are already starting to see that for example,

decisions made by some artificial intelligence systems are considered more valued, more truthful. however, this plays out the battle among big tech companies is on, and they're racing to develop the best i systems. what excites you about the future of artificial intelligence. let us know on youtube or shoot us an email. thanks for watching and have a good one. ah, a has to put did you do the full? i painted shannon testing. she survived. oh, shabbots things to music. and so he was the nazis favorite conductor to musicians under the swastika,

a documentary about the sounds of power and inspiring story about survival. music in nazi germany. watch now on youtube, d. w documentary lou time once again for a brain update. because this orchestra called the brain continuously adapts itself. and so we ask a few astute questions. are we smarter swarms or use psychopath? wouldn't cause it's monster waves. how powerful are your thoughts? we can control our thoughts, which makes us very powerful questions about life, the universe and the rest that we're series 40 to the answers almost everything this week.