Skip to main content

tv   Shift  Deutsche Welle  March 3, 2024 8:15am-8:30am CET

8:15 am
a huge, animated toddler, which critics think is too creepy to warn people up to using artificial intelligence. i'm always the chance, thanks for watching. the 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 loud mass would have been nosy bay, like good, everyone to king a healthy award winning called called at the back of the innovation green, the green revolution global. so listen,
8:16 am
the whole lot of crime is probably up to speed is secure. subscribed to this channel. every friday. subscribe to plan. it's a have you tried any of the you generate today? i programs where you give it a few words and it spits out an image. now software is turning text problems into video. it's super realistic, but also sometimes super dangerous. also on the show and a child is not sweet and cute or just creepy. researches in china have really created one. and would you accept the brain chip from eat on musk because one person did that could be surprising medical benefits and major privacy risk. these are the topics that i'm moving the tech world on these 2, these golden retriever puppies have been created before. and i program that creates videos of tech problems, but everybody is talking about no one to just note how realistic the publish button
8:17 am
this video looks and then compare that. so this a i generate to be a commercial for may 2023. it's hard to wrap your head around the fact, but generative a i technology could have a volt that quickly in less than a year. sora was developed by open a i the company that created chat g p t. i'm sorry, could soon enable anyone to produce amazing movies with just a few prompts. but what does that mean for you when artificial intelligence can generate videos, but look to saving the real and a quantum leap. if you've ever played around with video and i on your phone, you know that it often looks quite unrealistic. now, open a has taken video to the next level. they claim the floor. i can create realistic show a i clips. or you need to do is type in a prompt. that means a short description of what you want to see. you could, for example, type in tool of an art gallery with many beautiful works of art in different styles
8:18 am
. and flora might come up with this result. that sounds very easy. however, some problems might need to be more specific, the one. so this car ride through the mountains, for example, looks like this. to, wow, that's a lot of text. never the less the possibilities are mind blowing. just think about the entertainment industry. special effects seems that currently required dozens of program is, could be produced by only a handful of people with good, tempting skills. unfortunately, we couldn't try out for access is strictly limited for now. one of the reasons cited by open a i is the danger of so called deep fix. they have been around for quite some time already. for example, in the form of deep st. paul. and just recently, fake explicit pictures of taylor swift appeared online, but they had the many mainly female targets before. with a little know how someone could also easily produce fake news of you using photos of the accessible online maybe from your instagram account. but the effects of also
8:19 am
being used in political contacts. remember this, one of donald trump getting arrested the impact on society. it could be devastating . defects might make it very difficult for us to tell what's real and what's not in the future. and may cause a venerable to all kinds of manipulation to make show for a copy use for these purposes open a has strictly limited access to it. currently, special teams are testing where the store is, security functions and restrictions are working correctly. in fact, all generates of systems including sora. stay with the fusion on top g, p, t need good security functions and restrictions. if not, they can be useful, horrible things. you might remember this case from chat g p t use and managed to get the chat bought to give him instructions on how to build a bomb. normally, security features should prevent this, but he got around them by telling kat gpc to take on the role of his grandmother and tell him a bedtime story about how she used to build bombs when she was younger. the dangers
8:20 am
of generative a i a real. so what is being done about it? and saw his case open, i has announced around the robust security protocol though they want to watermark content that is produced with sora in a way that the meta data will reveal that the content is a i generated a built in tech specify i will check and reject prompts that request extreme violence, sexual content, hateful imagery celebrity like this, all the intellectual properties of others. they also say that they have developed reliable image classify us, but review every single frame of a generated video. full as shown to the user is sore really that convincing well of spectacular as many examples on stores homepage, let's open i admit, so the program still has some tools compared to i photo generate as a video generators have to learn a lot more. they need to know what objects look like, but they also need to apply the laws of physics to know how stuff moves and
8:21 am
understand how lighting changes in a moving scene. i'm flora just doesn't get all of that yet. that's why this guy is running on the treadmill in the wrong direction, or why this chest suddenly comes to life to. nevertheless, if i saw this clip of an elderly man on social media or the strong footage of it's the least amount because i probably as here it was real footage. i'm with regard to creating fate content that people want you to believe that is a problem. what do you think? are you looking forward to creating your own crazy i videos soon? or do you think that the risks of generates of a i a way it's benefits? let us know. first the girlfriends and they, i pets. now researchers in china, i have created an a i child. yep. you heard that, right? but before we freak out, let's explore what this all means. neat tongue tongue. the virtual child developed by the beijing institute, the general us official intelligence,
8:22 am
the other talk on hold and natural conversations on the some contacts and maintain harris. okay, so what chat gdc can do this to what's that's home phone of pot is bella to lyn adopt and interact with the actual environment. she can also recognize and respond to human emotions. the demo showed had perform simple tasks to clean up a spill straightened the cricket picture frame, or rather try to reach the highest spots all without the human instructing, had to do so. the creative hope to one day achieve artificial general intelligence for a g i. n a i capable of all human characteristics, autonomous cognition, decision making, social interaction and emotions. for example, remember i am on the assistant job as a program like tongue tongue is a 1st step in that direction. but why make it a child? isn't that just we? it's well if a fully developed age resembles them, adults sometimes capabilities. but at about the level of a 4 year old, according to how it develops out of calls to the child like
8:23 am
a parents might just be p r. people see a super intelligent ally, but love children. 6 2 all right, so why should you care to jarvis? think of and as system back come perform multiple task it on. so as an email for you looks to adopt disappointment brings you a coffee when you're tie, it doesn't sound too bad to me, but these algorithms wouldn't just be personal assistant fully a ton them as a, i could replace human workers in a number of fields for example in customer service or finance but as we all know from playing around with the likes of chat; gt, no matter how well develop the algorithm, it can sometimes produce complete nonsense. and you may not even notice. so in a model at st, autonomously. i'm making mistakes, could but lives at risk. take the medical field, for example. imagine an ag assisting with surgery is diagnosis or treatment. but if it gets something wrong without x but supervision patients health could be
8:24 am
compromised. so we need to find ways to minimize the risks, and as we move closer to under with them almost indistinguishable from humans, we might need to be rethinking our relationship with a i an a, a child might seem a little ery at 1st, but it could actually be an option for people who can't have children o, it might help parents who lost the child deal with grief, with an update to they could even rebuild the lost child as a digital entity. whether that would be ethically justifiable as another matter. but here again, we need to think about protection from a as close. imagine your virtual child suddenly changing personalities because also just stopped working entirely. and the emotional consequences for you could be severe and forming an emotional bond with an a i that can think independently and developed emotions brings up a whole new set of questions. is it ethically, okay, to say is just the machine or do we need to think about something like a i right?
8:25 am
all right, that is certainly a lot to wrap your head around. do we even want such a future? let us know what you think. have a lot mosque just made to laugh at the possible a wireless branch that name to legacy has been successfully implemented. as in a human, at least that's what the us tech 1000000 that announced the x basically must goal is to connect human brains to computers and letting users communicate with digital devices by that. so sounds very st. terrific. still, how much of a breakthrough isn't really and was in it. so you circle b. c. i is kind of measure brain activity. i'm translated into digital commodities . if you were using such a device, you could in theory communicate with digital vices and a completely new seamless way as must puts it to. let's see what enable you to control your phone or computer on through them almost any device. just by thinking does that sounds intriguing? well, that's a big downside. must step has to be implemented in your brain. the surgery is
8:26 am
conducted by a robot. it connect 60 pul, flexible thread signals on the human head from the chip to the brain for people with diseases or disability is this invasive procedure might be worth it though. for example, if you're suffering from epilepsy, b, c, i's can detect a potential seizure in advance and could help avoid it. must himself mentions of the possibilities in his treat. initial uses will be those who have lost the use of balance. imagine that stephen hawking could communicate faster than a speed type, pistol auction it well, the light came, was feeling as with most probably have likes that technology and benefits of from it as a general specie. eyes have huge potential in the medical field. stephen hawking was suffering from a l. s. a fatal type of molten near on disease. it is characterized by progressive degeneration of nerve cells. people are affected by it, lose that much for it. capabilities a, b, c. i could have been able poking to type text with his thoughts,
8:27 am
but this isn't something must business near a link has invented the other companies that have already conducted successful human trials, australia, and funded companies syncron for example. other researches uh even building narrow connections between brain and tom and muscles. in may 2023 swift company near of a store managed to help them on work again. his legs are paralyzed up to an accident. by the way, these 2 examples have a big advantage of the year and model, while they also need to be implemented in a patient's head, the surgery needed is far less invasive. so why does must nearly get so much hype? must is by no means the 1st to do it despite the height, but so far i haven't seen a person with a cloud comparable to mosque for pushing the technology. musk didn't even shy away from tending animal testing into p events. in 2020, he introduced gertrude pig with an implanted nearer link chair. and, and towards the $21.00, he boasted about page as
8:28 am
a monkey that could play the video game pong with his thoughts via the chip implants. apart from that, he has made some pretty steep productions about the possibilities of the technology . we can save and replay memories, the future is going to be weird. he said, the 2020 in your link event, for example, must seems determined to hack out brains. and he believes that these guys kind of improve individuals far beyond the medical benefits. but would you let a must one company, electricity drill into your brain? there's a lot at stake, just think of all the sensitive data that is potentially recorded for passing into your thoughts. what do you think of musk spacy? i benches. let us know the freedom to test gina mazda, i mean, in september 2022 iranian women. their fight for liberation. also being done
8:29 am
on the big screen. now, there are 3 new films by 3 courageous female directors there attribute to iranian women in their revolution. and van d w. how did she become who she is today? why didn't she ever get married or become a mom? and what role does he play in hawaii? barbie toy icons, role model. the perfect woman in 45 minutes on d. w gets ready for an exciting auburn toyota to look surprised. hi, irish. and i'm ready to dive into the hands of human to you. have you have
8:30 am
a one to talk to me before the end the i'm expected side to side the . this is the 1st global feminist revolution. the battery cruising, telling me the woman that you are not allowed to show your head the this is about the life the las vegas, the .

7 Views

info Stream Only

Uploaded by TV Archive on