Skip to main content

tv   Tomorrow Today  Deutsche Welle  June 23, 2024 11:30pm-12:01am CEST

11:30 pm
a little surprising i was shoving and i'm ready to dive into the hands of children . to us. you have you have a one dodge, so do you need a quote on that? we've got a response and the i'm expected size of the optimist. few artificial intelligence as the solution for practically everything. how will we take care of global crises in the future with a couple of cars navigate with our drivers with a i probably finally secure will piece with they are. but who should be held responsible if the icons is damaged or injury software, the user or the manufacturer? this time around would be looking at that question and many more post by the field of artificial intelligence. welcome to tomorrow. today.
11:31 pm
he's nice to artificial intelligence will change our lives in fundamental way. but who should control the development of this powerful technology? should it be driven by large corporations? or can it also be made available to any developer or user as a public good? the technical university of munich as an important center for a i development and research to, to mention a give to our university is considered a leading university in europe. insight in particular, we're at the head of the pack in the fields of robotics and machine learning and perception of fields along with around 50 other universities and companies. the t u . m is involved in the a alliance of project, initiated by the meta and ibm groups for the development of artificial intelligence . institutions like southern are also members there, photos, coast, ocean, close all lab ration between research industry and says,
11:32 pm
site is crucial. this means that we want to work together to advance responsible technology as long as you know, but we also want to make them transparent and understandable and available. and i would say a very important contribution from industry is to make resources available, but also expertise wants to sort of, especially in a i where many leading lines are also anchored in industry to industry for on good . some of those leading mines, however, are not usually associated with serving the public interest. how much real influence, for instance, will matter, which also owns facebook and instagram have on the a alliance. how trustworthy you're such companies when they say they're committed to transparency and the common good members of the lie on association are critical of such an alliance. a decentralized network of scientists, i t experts and lead people lie on advocates for open access to artificial intelligence. the community views itself is not turn it up to the commercialization
11:33 pm
of a guy and doesn't want to lead this field to big tech alone. but i own an estate for a large lie on stands for large scale a i open network. and the idea behind it is to create an open source community of people who want open source a i, as in something like an open source chat, g p t that anyone can download. this one is that we bring these people together to work on free, open, transparent, a i, it's all in the associations founders fear that companies driven by commercial interests could simply revise the promises made by the alliance when they get in the way of higher profits. so is there a danger that developments financed with tax payer money might be used by large corporations for commercial purposes? somebody had a dean doesn't think so. the end of this talk is at the end of the day,
11:34 pm
of course we also want to ensure that the things that are developed there are not all lovely together and simply exploit and uncomfortable. but i think there is cooperation at the table and try to cope with one of home, who adults with home, despite the reservations, the challenges posed by a development or so shoot that no one company can tackle that alone. basic research as universities need to expertise from firms to develop market ready products. as a, a twin, the door that we're doing here means that globally, many of the most important players have recognized that we must do this together. i think we need to work together to create the foundations for this technology to become available. and then it should ultimately reach the real world to buy a product, to services and above all, and via open source. and that's a driving concept, and it's very important to us. the open source is also the foundation for the lie on community which has already developed very successful products based on it. and
11:35 pm
by state one example of how the whole thing can work as our lee o l m, which is a german open source model just like chat g b t, you can download it. and the more advanced version of it is just as good as chatting detail, and it's even better at translating between english and german. well, that's the 1st thing, this annoyed but transparency won't solve every problem with the technology to copyright issues. for instance, recently the new york times and several authors filed a lawsuit against open a i, because it's chat box use their texts to train artificial intelligence. my on was also sued because the sources in the image database is it uses are always clearly marked. we understand i'm for who we just put the length is which are listed in common across anyway as a. so in this non profit research database on end, we make the data that belongs to a successful to the public which provide. it was always in the public domain, like filtering it out was a pain on so we just make it less annoying and easy to check. that means some
11:36 pm
people pictured in the database is where identifiable? many legal issues are still unresolved. this is slowing down development for both corporations and non profit organizations. a, like somebody had a dean and coast of schuman, i agree on one important point of desktop, caught in the state, in germany here in that area. but especially in europe has clearly invested to level 2. it was a status in this. it's a shame that countries didn't recognize this much earlier and say we're investing so that this technology will be developed by us as in other words, that kind of apollo program for a high level to help us all to work better live longer and be healthy or not so that would be so great, but states haven't done that money starting the europe doesn't want to fault even farther behind in the race. it will have to invest more in the development of
11:37 pm
artificial intelligence. at the end of 2022, the company opened a i launched, shot to be an interactive chat. but the answers questions using knowledge harvested from the internet. within a week the platform had a 1000000 users. but what has the a i tool really done for everyday life? positive, for instance, actually changed learning and working at schools and universities. hello pine. the nice to see you. this, but i'm not actually real, but you probably already knew that phone location. there's probably multiple versions that i mean now because you chat g ctx that seems to be so kind as to become a hot topic if that's all each world. if i was also created by computer software, finish destiny. so i think the intelligence development of animated avatars like
11:38 pm
these has been optimized by a i systems like chat t p t. but it hasn't changed to learning dramatically yet at least not for the students at this school in hamburg. whitney says connected, and that is when i do use it, it's more for research to understand a particular topic better. so i conducted on a task better. instead of using google like my, just to use the chat cvt as a research tools to understand a top that in a p p t sometimes writes pretty strangely. so i just look for stuff using normal search engines, myself. smart. yeah, i like to be responsible for my own work to have control over is what i need. so it's just not that relevant for me. i thought i think it's a good one for me. after chat, t p t was launched, kids did try to use it at school and to do homework. but teachers have sense learned how to recognize that by asking what is the student really capable of and where did a i perhaps help. that's us knowing the civic and novelty has worn off,
11:39 pm
that we have to educate the students to use it consciously. palm is a worth to mrs. put the emphasis on words and not on tools them as and because we all like convenience, dependence, appealing to get results quickly on to something that doesn't help us have signed the heights. i would have to say, a fast overview of a topic, a new recipe for a request for a free retail. the tech's to offer as plausible responses to pretty much any demand, but is that intelligence? the technology is making us ponder that question. as you all decide, people shift to whom because it's brought the issue to the for because although many people have been using a i all along, they don't see it as artificial intelligence system. again, such theme unlocking your cell phones using search engines, looking for recommendation. all that's the only i to me, but it isn't in toronto for the child thought that gives the impression that it's, i think, and what convict us to it's a thinking and, and again,
11:40 pm
it's probably the most important question at the moment as what a, i should be allowed to do chat g p t. you can generate porn content that glorifies violence, fake news, fake websites, this information campaigns. ethical filters can identify and reject some of it, but software like that can recognize everything. the explosive growth of a i, applications is making even some insight or is it i t companies nervous development has been compared to pen down x or nuclear war them and even see to desktop. it's when you see that it can be used to train things like drones and other devices that are then allowed to search for targets on their own. not really serious like worries. of course, we also need international agreements here. so here for cause i splish, an upcoming. most research in the field focuses on assisting people and making work easier, but a, i can be used for much more to address other major problems. for example, like those involving military conflicts with researchers are looking at for
11:41 pm
artificial intelligence for the un to support it's blue helmet peacekeepers. they're trying to identify signs that indicate when a violent outbreak is imminent, the more violence displacement the people of the democratic republic of the congo, know them all too well. the un mission to the conflict region is supposed to help stabilize it. what if artificial intelligence could recognize the signs of impending outbreaks? providing peacekeeping forces with key support. that's what minutes, how the saudi is working on it. this was federal institute of technology or e t h in zurich. her teen feeds data on the countries past complex into a i software pose. it's one of these in best when the car, the not we can see what their cord notes are. so weird they happened. and we also know when they have those with us here. so we can then see how the conflicts
11:42 pm
roadside conflict. that's like the samsung da i process is thousands of pieces of information including reports from the media or un peacekeepers that's enabled the research or to retroactively identify sign certain violence outbreaks were impending midtown by mountain awesome. events, you can recognize patterns, part of depends on how systematically the data is recorded and its quality clinical entity fee. so how systematically these conflicts also produce data to do to you via email, is with a ies help to recognize the signs of impending conflicts. that would be very useful information for the u. n's, 90000 blue helmets worldwide. elizabeth, the organization's peacekeeping missions to have proven effective, says security policy expert under the us bank. once the task complete, we can see that conflict hotspots spread less quickly and widely house pride. and also that the difficult transition from more to peace can be managed more swiftly.
11:43 pm
that would be a bad think they have to come. and of course, the better the information, the more forward looking at was the better the un peacekeepers will be able to work and allow hating the are by that's why this with institutes president has come to new york to steal a cooperation agreement with the un safety of 70 of them today will cement this cooperation at an institutional level. the un doesn't often make such a great. so this is a huge opportunity for the new ca shifted, but also for switzerland scribes. the if you have few bites of design, t h has a lot to contribute really by combining strange and in computer science and social sciences zions to help us recognize the conflicts and potentials earlier. that's a lot faster and provide more targeted support for you to i can sion and it's a 100 and there's an incredible amount of potential this owncloud place for you to put inside jeanetta
11:44 pm
t v. i as in almost potential for goods in the village scale. not the least artificial intelligence can and must also serve peace and security of the un security council, which is responsible for world peace, dedicated a session to ai swiss representative. and ambassador, a postcard about his views says it's vital for the un, huge data set to be evaluated more systematically. however, really cool. this may, i might help to predict developments faster and more accurately, but it can never replace the need for people to take responsibility. ultimately, human dignity must be at the heart of these efforts, and the international community country's interest is un and the security council must take responsibility in order to find sustainable people until to give them the food. i know how to give late the theme to the un can often do little though, like in the current situation in the middle east, in the future could a, i recognize when
11:45 pm
a war is set to break out creek sounds pre k live and see it or outbreaks remain very difficult events to predict because fortunately they are very rare events on independence can be better predicted. the more frequently they occur. comments before, what we focus on are very localized conflict dynamics over a very short period of time. so we have more information getting to them on and artificial intelligence systems can make a contribution to the game, especially when they're combined with human expertise causing it st. meet names like you'd expect the ser, algorithms that can promote peace are still in their infancy and training them on solid data is an ongoing process. but in the future, humans and a i will work together to help identify potential conflicts before they occur. so it's no longer in its infancy autonomy is driving, isn't all grown up just yet. functions like emergency breaking,
11:46 pm
keeping the car and its lane, or other driver assist might have grown quite common. but computers still aren't quite ready to take the wheel completely. the it's already being showcased and promotional videos ended motor shows. the dream of self driving cars. just climb in and let artificial intelligence chauffeur, you wherever you want to go. but the systems weren't that far along yet. gimme a follow up and put in a more driving task. so not tournaments vehicle takes on the more so called corner cases. you get is that those are quite rare occurrences that also must be tested for and saw what's challenging is there are so you radically an infinite number of them you to get expert young back doesn't see unrestricted, fully autonomous driving for at least another decade. still boxes and what are already being sold and will grow increasingly, comments or assist systems and highly automated drive intensive team. there i can
11:47 pm
let myself be driven on certain sections of the route, mike and stuff and go traffic on the way to work. go for k often, vic, it's all but such vehicles have already had german streets. mercedes was the 1st domestic car maker to launch a highly automated driving system in 2022. the drive pilot at least on the highway and it speeds under 60 kilometers and our drivers can take their hands off the steering wheel and turn their attention to other things. such technology always raises the question, who is ultimately responsible if an accident occurs like and this crash involving a semi autonomous vehicle in taiwan. legal experts are trying to answer that questionnaire and others at germany's highest transport court. find all to know with a ton of mist dr. h. faulty products rather than driver. an error will be the prime cause of accidents. so manufacturer liability for drivers will increasingly become
11:48 pm
a topic of interest and then for everyone to have to. so legal experts are looking for ways to more justly a portion responsibility. one goal is to make it easier to prove when a guy is actually to blame for an accident. there are currently no fully automated vehicles driving on public roads anywhere in the world . there is one, however, that is allowed to drive autonomy as we in germany. the only on fields artificial intelligence looks at to revolutionize agriculture as well. the edit test farm in the german state in saxony and event showcasing the agricultural machinery of the future is about to begin. machines like these are designed to work the fields a ton of misleading. so without humans on board. some are powered by solar energy
11:49 pm
and can turn on a dime. most are equipped with sensors, lasers cameras and gps receivers, and are controlled by computers. coastal miller is competing in the event with a robotic vehicle that also generates all its own energy from a solar away the farm droid. as an economist reading robot, it performs the task without the help of chemicals and is therefore eco friendly. far enjoyed follows. rows it made in the field using a set of small but short blades. it cuts down every plants that it didn't, so itself relieving with the machine it with the machine. we plan, see the pre determined position was and then no later. ok, we planted some things here, so we want to whole around it. that means we can hope before the plant is even visible. it's on the scene is farms royd remembers every spot and so to see and
11:50 pm
who's around the travels around the field at less than one kilometer per hour. the solid dollars is this at all takes time. it's slow. but that also gives us a very high level of efficiency. if it's cut, it has 24 hours because fund royd works around the clock. it has to because it can't get we instead of growing too high. it also only works on flat fields through soil isn't too clumping because it moves so slowly. however, farm toward is so far, the only fully economist farming vehicle to be allowed out on german feels. as a months it same years. you can see that it stops, might have in calendar is an obstacle and much more just as something gets in the way whether it's a person, an animal or an object. so the machine is say, it's the most, you know, a few years ago, a tractor manufacturer fund caused quite
11:51 pm
a stir with a vehicle design called a salvo. the plan was to use a fleet of small autonomy networks, tractors, instead of a single large one. farmers would supposedly only have to monitor them from a distance, but the project didn't work out in part because wireless data networks in the country side simply weren't powerful enough. time for the agricultural robots at the event to show off what they can do. their designers make a final check the the field has been precisely measured using electronics, and the data has been digitize. the digital map that is bought, it has been transmitted wirelessly to the robots, the farm droid weed, or is the 1st machine to strut at stuff. they weighs less than a ton,
11:52 pm
so it doesn't compact to soil much. the need to be fixed as a side effect. you reduce load on this oil because these machines aren't as heavy as their content for eric, how many flight hooked up and a 300 horsepower machine with the tractor and mounted implements are going to weigh at least 20 tons so that i'm pulling 20 tons across the field funds each time and even luckily for engineer thomas had did see is that's another major advantage of agricultural robots that can be designed to be lightweight. what is the state mom, the guns and bang and put in? so products systems have a lot of potential for a more flexible use with less wage with which leads to significantly less soil compaction. and that ultimately pays off in terms of yield, soil life and overall field health and our office hours. the ad pod robot tops the scale that are relatively light, 7 tons, and it's caterpillar tracts. distribute weight better. that also protects the soil
11:53 pm
. the light flexible maneuverable, more or less automatic or at times even autonomous controlled by humans. but none behind the wheel. the future of agriculture is already a reality on this german field. one that will help make up for a shortage of workers on the countries farms. the development is read. why do you have a science question then send it to us as a video, text or voice message? if we answer it on the show, we'll send you a little surprises the thank you. so come on. just ask this viewer question. comes from over a in governor how long can days will be stored?
11:54 pm
humans have recorded their languages and their knowledge in the form of drawings and symbols, since prehistoric times on the on log data carriers from back then play tablets or stones, much of the information. so by, for thousands of years, as a result, these ancient data carriers can be read and decoded. today, the modern storage devices can carry a lot more data, even though they're a lot smaller. the life span of most cds and dvds is estimated less than 100 years . the data that's stored on the kind of huge server is operated by large companies, should theoretically last forever as long as the days is repeatedly copied. and as long as defective hard drives are replaced in order to
11:55 pm
keep the ever increasing amounts of data under control, researchers are developing smaller and smaller storage devices. in the future, synthetic dna could replace conventional data carriers. these artificially generation genetic strands would have more than a 1000000 times the capacity of modern hard drives. and the reconstruction of the onto the whole dna has shown the dna molecules are still legible, after thousands of years, or another method of storing data in courts. despite the technology uses namo lasers to write information into 5 dimensional fused courts of mind boggling speed . these glasses, the size of coins can store up to 360 terabytes and they last a very long time. up to 13800000000 years. whether or not there's someone in 13800000000 years who can read the data is another
11:56 pm
question. and that wraps up our overview of artificial intelligence here on t w. science show. thanks for joining us. thank you again soon, but tomorrow today the
11:57 pm
are you a fan or are you are a real fan for the years in germany this summer we want to know what makes german football culture so special the you just have to see to go the coming of age i'm 50 down this is yeah,
11:58 pm
i think it was and someone came in the opposite. advise them about that. how much freedom is on in this tiny house constance with conversations, cooking and music. oh, it's on van d w. the you might see these how much can we do simultaneously? multitasking these, the modern methods. because if we do too much at $180.00 all wrong, we mess things up, risking brain damage. so let's stop this self sabotage,
11:59 pm
humans and multitasking watch. now on youtube, v w documentary, this shadows, these pop costs and video shed lights on the dog is devastating colonial har is infected by germany across and he employed scorched most tactic farms and destroy livestock. what is the legacy of this wide spread race as depression today? history, the just, we need to talk about here the stories, shadows of german colonialism. the dream of revolution in 1979, the hate to dictate is the most that was forced out of nicaragua. the people hoped for sarah society and received international solidarity. i imagine we would change the world tens of thousands of volunteers to help with reconstruction.
12:00 am
but the end the dream of revolution, july 20th on d, w, the . this is dw nero's, and based on the top of the story is ro says it's investigating. the conduct of the soldiers off to the video was really is showing an injured palestinian strapped to the hood of an omelet car in the northern westbank. the man's family says he was struck by his very gunfire near janine, the un special patrol accused of fail using a human shield. russian said, media report that gunman have a tech to send the gold and to orthodox churches in the southern republic of august
12:01 am
. done.

11 Views

info Stream Only

Uploaded by TV Archive on