tv Shift Deutsche Welle February 3, 2024 6:15pm-6:30pm CET
6:15 pm
are attached so septic floors, how artificial intelligence is being used more and more to both good and bad aspect . i'm really thank so much for your company. the . when the started we started to understand clearly and what you need to do for us. ukraine was assembled. are we moving towards europe or not? the 2013, the still is afraid democracy protests in ukraine. the country response to your price to ship to west have recognized the danger and the terms of being foretold on the signs with us. we really understand that
6:16 pm
so easily for you. uh, we will lose our concepts, euro my done jobs february 17th on d, w. imagine you have to write a law or a basis, that's a lot of work, but then you want to kind of a i do it better than icon both on the show. why is one of the world's biggest future best hunting to x? formally twitter post as the videos does even last one to make x when you take talk and it's a disturbing phenomenon. but a i generated nude picks us searching online, and kids are using them to bleach of that, even though it is illegal. these are the topics that i'm moving the tech love with you soon have to turn to x to what's your favorite creative in january, 1 of the, well, the biggest need to but who missed the beast, dropped the video and x for the 1st time instead of inches and this is happening as
6:17 pm
some big creat says, leaving platforms like youtube. so can you loan most leo, them to ex all set and sweat the tank very enough time. so turn it into an everything up with a focus on videos to do that. he needs creators to actually post videos to his platforms. until recently mr. based wasn't too sure about that idea. he said his video is cost more to make some what he's on the platform, but he'd be willing to test it once the monetization is cranking. and apparently it's cranking now, his 1st video and ex pulled in over a 160000000 views, bring in roughly 230000 yards. however, not everyone is convinced that this happened organically. some skeptics think must pull some strings behind the scenes, pushing the video into everyone's feet, even less the based himself that advertises bought more ads to the video than they might have done otherwise, suspecting the video was going to be big. however,
6:18 pm
bringing popular content creators to ex is still a better attempt to make it a video platform than what most tried so far. last year he allowed us, as opposed to our videos, leading to pirates and movies popping up full of x foot. well, what am i need to do to create the perfect video platform? let's check out what the competition is doing right. and what they struggle with. use of love, short, snappy verticals, such as a tick tock is the app to turn to when searching for news tutorials, products with news or recipes. the rise of textbook other platforms have tried to copy it success. and that gave us instagram reels and youtube shows both worked. at the end of 2023, maybe 75 percent of the most. what you tube channels are mostly producing youtube schultz don't. instagram reels, received a lot of use each day, spend regular photo pulse. so vertical seems to be the way to go fax, but creative,
6:19 pm
i'm making little money with them. ready the problem with short videos is that they don't make much money. the reason for that is simple. the longer the video, the more ads platforms kind of crowd into it. and that's why well, all the other platforms i'm pushing the short formats take, doc is actually encouraging creators to post longer videos. now up to 10 minutes, for comparison crate is in between $1.06 for a 1000 views on a tick tock or youtube show, but they can make between one and 29 year old for a 1000 use on a long youtube video. this is according to create those who shed their revenue. you could argue that x should still just focus on verticals because that's what the views up, but keep in mind can create as well and usually leads to more high quality content for us to enjoy. okay, hold up. so ex, should be focusing on long videos off the old, or is that another format that's possible? ext, breaking into streaming live streams booms during the pandemic and youtube instagram
6:20 pm
and take talk all adapted to this trend by improving the live services. musk seems to have noticed that as well, and recently hosted the 1st live stream on it on himself playing video games. it could be a strategic time to recruit stream is to most of the a series of controversial decisions by the type of streaming telephone switch stream is looking for new options. some of the biggest, which stream is have reportedly been offered, unaccepted millions of years. but other platforms stream exclusively on that and it says one thing, but musk doesn't black is much just the warning we will talk about a i generated fake news. images of mine is some of the you may find this to study a i generate to child abuse, images law searching online. sometime ago teenagers in the u. k. and spain use they, i to create explicit pictures of them is most likely to believe them. how is it possible to generate defects like that in the 1st place? and what can you do to protect your child for yourself?
6:21 pm
exploit to, to the dfcs. first it was, celebrities became victims of pulling deep fakes. but with the technology being widely available and becoming easier, today's, the problem is spreading. studies have shown the 90 percent of victims of women. that's a side statistic, but kind of predictable. so what happened exactly in those british and spanish schools? well, you case schools have reported that students are using a defect tech to bully that peers and dozens of spanish teenagers said they received a i generated naked images of themselves on their mobile phones. the pictures were created using photo showing the targeted kids fully closed. many of the picks was sauced from the girls social media accounts, and i program then generated a virtual image of the person without clothes on police identified at least 11 local boys who we, they created a set can a to the images by once up until the ground incidents like these are not limited to spain or the u. k. they can happen anywhere in the world to adults and kids like
6:22 pm
that's disturbing on so many levels. it's a crime. the kids from spain in the u. k. might not have been aware of this, but creating and publishing explicit images of other minors is a crime in the countries whether they're using ai or not. legally speaking, we're talking about child sexual abuse materials making, possessing, distributing such material is illegal in many countries. the fact that the images aren't real doesn't necessarily change the fact that it's a punishable offense. for example, in another case in south korea, a menus artificial intelligence to generate more than $300.00 exploitative images of children. in september 2023, it was sentenced to 2 and a half years in jail, even the cartoons or other less realistic depictions are illegal in many countries . it's easy to do. it's a lot to all of a few clicks to create very convincing photo realistic content spot us teenagers use an application which promises to undress any go using i. it emphasizes that
6:23 pm
it's guidelines include age, verification and obtaining consent, but it doesn't stick to the it's a fast growing problem. such materials threatens to overwhelm the internet. that's according to you, kid based, non profit internet. what? foundation? 3 reasons why this is a huge problem. firstly, a i generated images of child sexual abuse, couldn't normalize babies of real children. secondly, it could make please what call the law enforcement offices could find it more difficult to identify whether real children are risk. and thirdly, i generated images of child abuse, could hyphen the risk for children in real life by encouraging peter files of facilitating greeting. what can you do to protect your child and just felt bad news 1st, as the spanish case proves any photo can be used to generate deep fake, pornographic images, b, o, y, and inform your child that this kind of technology does exist. prepare yourself and your child to be faced with that kind of material. on a case that happens,
6:24 pm
inform the police, the person sharing such images is carpeting a crime and many places in the world, no matter if the material is generated or real. i think we also need more regulation on generative ai companies that offer a i image. generators should have said gods in place to make it impossible to create virtual needs of someone without that consent. in october 2023, a low written by chat j b t went into effect in brazil without changes. it's not a good thing. and a university in prague in the czech republic has deemed writing about slip basis points us. that's because they can't tell if students are using a i to write it. what does it mean if i can write as well or better than us? brazil, a i writes little text pads should not be charged to replace stolen war to meet those. that's the basic just of a new law put into place and the resilience that d a pull to like right? so far, nothing special. but get this. the draft that became law was 100 percent
6:25 pm
a i generated a council man. how does it in, without informing his colleagues that the proposal was written by chat g p 2, it was put into a act without any changes pulled to a library's man, only found out about the laws real origin by social media. when the accounts of mine bragged about his achievements and the already of the examples from around the globe by lawmakers use the i in massachusetts usa, a democratic senate to attend to chat gt to help write a bill aimed at regulating i, including chat gpc, could i write battles, little x, but at least seats attentional chat, gpc maple 10, then even more of a mental shift, then the advent of the into that space, the renowned harvard law school in the us. but have it also warns of as potential shortcomings. g b t may not fully comprehend the nuances of the law. this could get dangerous in situations that required in that legal analysis. but it's popular one year after
6:26 pm
it's released chat, gpc had more than a $118000000.00 monthly active uses. and is one of the fastest growing consume apps in history? it looks like artificial intelligence has already changed the way many of us right . not just in regard to legal text but also academia a i writing a bunch of pieces. you google bachelor pieces with chat; gt; you'll find a lot of information on how you can use chat, j, b t and other a i tools to write the assign to the paper by the way, i'm not saying you should do it. some universities have reacted with, complete a i bonds, but unfortunately it's very hard to identify which techs have been written by a i. as in july 2023 opened a i, the mother company of child cvt shut down. it's a, i classify it. so it was supposed to determine the likelihood that a text passage has been written by another guy, but it has an extremely low rate of accuracy and got it wrong 3 times out of full.
6:27 pm
other universities have called for most student presentations instead of writing to get around the problem. remember the university and probably mentioned it completely, a bullet spot so that the sees the new students. that's the 1st move of its kind of plagiarism has always been a problem in science. now a chat box have made it even worse. the shortcomings of a, i think large language models visa. i programs the process and generate human like language they kind of produce and a thing that is not in that database. so they won't come up with new ideas or solutions, but more or less copy paste parts of the text that they've been set. and there's another problem, hello and what like very sophisticated or the complete tools they repeatedly try to guess the next would and a sentence in the process. they are prone to making a false information, a phenomenon, sometimes called hallucinations. that means all chatbox introduce false information . this ranges from about 3 percent in the most advanced jpg model to
6:28 pm
a rate of around 27 percent for one of google's models. wrong information. any of this is my main. you won't get a good grade, but it's a little based on a wrong information. might negatively impact millions of people. however, i can still take a positive role here. it can help with some of the more tedious elements of law making. for example, searching and citing laws that already exist this good significantly shortened the process. the politician inputs the great claims, but keeping the proposal origin a secret was intentional and but his objective was not just to resolve a local issue, but also to spark a debate on the potential benefits of a i that's all from me today. but i'm see you next time the have you ever heard the term jump? ha. just basically describes how masses of young people are leaving the country to
6:29 pm
search for better opportunities elsewhere. when field work is leave africa, they leave a gap program, a great, great big. it's hard to appreciate the real part of this problem. we sent out our team of correspondence to find some answers this 77 percent next on d w. but you know, here we are happy that way of boxing the story. we have a getting a visa is more difficult than finding gold hosted to use the force and for the future in the stories industries that are being discussed across the country. news africa. in 60 minutes on d, w. c, credit slide, we discovered new adventures in $362.00
6:30 pm
and explore fascinating. both heritage selling dw world heritage 360 now the to have you ever heard the term job palm? well, it's quite a buzzword in nigeria, and it basically describes how mass of young people are picking up and leaving the country to search for better opportunities elsewhere. we met some of these people and prepared bill stories here for you at the 77 percent live show and i'm your host to date. the here's what's coming up in a board. yeah, we need an id expert who has said his eyes on a job in canada.
8 Views
Uploaded by TV Archive on