Skip to main content

tv   Shift  Deutsche Welle  November 1, 2020 10:15am-10:31am CET

10:15 am
think of a country you know where that's whiskey or a beautiful scenery i'm sure cordie was definitely also one of the woods. with the death of surf sean connery the movie world has lost a legend. her room is strong. and certainly has you watching day w. news up next did every technology program shift with a look at using facial recognition against covert 9 tane i mean like information on any of our stories you can always visit our website www dot com i'm rebecca reaches in bed and thanks very much watching. every day. for us and for our play of. the ideas is on its way to bring you more conservation plays how do we make cities cleaner how can we protect animals and their habitats what to do with
10:16 am
the waste. we can make a difference by choosing reforestation over a diesel or station recycling or disposal smart new solutions oberstein said in our earth is truly unique and we know that their uniqueness is what allows us to live and survive the ideas the environmental suit through global 3000 on g.w. and. faces of the fingerprints off today thanks to facial recognition technology some use it to unlock their phones convenient for others it's used to surveil their every step what does face recognition capable off at how can you escape it topic today on shift.
10:17 am
$64.00 countries around the world use facial recognition technology the hardware and software is often produced in china low income countries can even get reduced interest on their chinese loans that's what the carnegie endowment for international peace found out not much is known about the global industry for facial recognition technology they act discreetly even if some well known tech companies them and the russian capital moscow facial recognition technology has been recently implemented to the 1st stress test the corona pandemic. more than 100000 facial recognition cameras are being used to fight the spread of corona virus in moscow. the cameras were allegedly monitoring chinese people in a targeted manner and they've later been used to track whether people in quarantine were really staying at home. activists see facial recognition in geo tracking as a threat to human rights. or some measures can be justified
10:18 am
but it's up to us to make sure that after the epidemic ends these measures are lifted. within weeks software companies around the world adapted their systems to the new reality under coronavirus measures facial recognition now also works with face masks. when people don't wear a mask the recognition rate is about 99.5 percent when wearing a mask the recognition rate can still reach about 95 percent. if the cameras connected to a temperature sensor it can even measure a person's body temperature and inform the authorities if someone has a fever. in order to recognize humans wearing masks the systems need to practice using photos companies have repeatedly used photos from social media for this so called scrapers software automatically downloads all photos that have not been set
10:19 am
to private the new york times revealed that the us company clearview collected brace yourselves 3000000000 photos off twitter facebook and co the resulting database is used by us police forces to convict criminals but how does facial recognition technology actually work here are the basics in 50 seconds. most basic recognition systems use a 2 d. source like a photo or video and algorithm analyzes the relative position distance and location of basal features such as the eyes nose and mouth. alternatively the algorithm analyzes multiple images of a specific person and generates a data version of the face for comparison. 3 d. facial recognition technology uses sensors to capture information about the shape of a face like the exact depth of someone's eye sockets. texture
10:20 am
analysis uses details of the skin such as unique lines patterns and spots to identify a face. how secure is facial recognition technology and how far could it potentially reach we asked me a check from taiwan she is a data scientist and expert for machine learning. the morning person that they're going to get this database is more about level for the hackers so you can think of like once you add to deface image to your data then this kind of like if you want to hack in it because once they have your face they could kind of pretend they are you. maybe you were 1834 i don't know 87 different points on your face so that say if you cover your nose and your mouth so you kind of covered to certain numbers of the
10:21 am
data point of it so then the rest in number that you could use rippy that they have point to from your eye brows from your eyes that distance between your 2 eyes and your cheek bone so still you could get certain of the features of it but take a bit longer time to compare of it even some people will wear a t. shirt with a human face in the front so that the camera is kind of confused whether this teacher is year or does face is real you. there are some plus in the model in between so i think we definitely need to think about when we talk about the promising we also have to think about the panicky why they do the things run their experiments or a survey from research center they use 500 people's face of america. and then they come paired it with the criminal to the set and then they're found it is kind of like $35.00 case of it are match with
10:22 am
a common data but they are not in germany we usually discuss data security and privacy when we talk about facial recognition that's justified but do we also see the positive side in india this technology has reunited thousands of missing children with their families and in china too many people rather through the process but. when menkin young goes to eat lunch high school woman hang joe all she needs to show is her face and she receives her preordered with bacon vegetables she no longer needs to queue to pay. than it used to be we'd always queue for ages because everyone paid by card students can also borrow books from the library in this way. tells us exactly who has borrowed which books and one this helps the library to cater to the students interest when buying new books.
10:23 am
some of the plain jane young has gotten used to the cameras. on the one hand she do you sort of get the feeling you're being watched on the other hand it does help us learn and that's the main reason we're here. for all the pros outweigh the cons. so. screens are used to display when the boarding school students are in the rooms on time in the evenings they also show who's late to class in the mornings modern tech. knology is supposed to help the students excel says the school the more data the better. and san francisco things are very different there facial recognition was officially prohibited for use by the city's authorities in 29001 paskin was the initiator of the built in psychologically unhealthy when people know they were being watching every aspect of the public row on the streets in parks that's not the kind of
10:24 am
city i want to live one reason for the ban on this technology fear of racial profiling indeed facial recognition systems make more mistakes with asian and black persons as well as with women children and elderly people study by the mit named gender shades shows that the algorithms work especially bad on black women in response to this study i.b.m. brought out a new dataset called diversity in faces it consists of creative commons photos of 1000000 people of different skin colors and can be used to train data by the global research community in mid june companies like i.b.m. amazon and microsoft officially declared that they would not supply face recognition technology to the police anymore but nowadays software developers are not only working on artificial intelligence to recognize our faces but also the mood where. human emotions are complex as is expressing them using the face
10:25 am
scientists at the german found institute for integrated circuits trained algorithms to recognize emotional patterns in human faces but universal expressions even exist for emotions like anger grief and fear. since different aspects of the things express emotion and certain muscle movements create a look of joy anger grief and so on from that's universally understood in most cultures and the things we can discern from the facial expressions of another human are things. that we can also teach software with a very high success rate. to train the software the research is also use photos of faces with clearly defined expressions that system for analyzing emotions can detect anger joy rage and grief in real time but what can it be used for. and the. typical thing that emotion ai is used for is interactions between
10:26 am
humans and machines and one of our projects is teaching a robot to work with autistic children to help them read non-verbal signals and facial expressions. and that was angry well done you did that well. the technology is also used in the automotive industry here it's primarily used to monitor the driver's emotional state and spotting signs of distraction or stress. so is a motion ai home that's. these kinds of technologies have the potential to be used to monitor and manipulate people of course that's precisely why it's important to actively engage with this technology and its potential consequences we can only do that if we are in control and really understand these technologies and what they can and cannot do with them and. it's becoming increasingly harder to escape facial recognition but one can still try to trick the algorithms
10:27 am
some creative ideas on how to do this have been developed in art and design. this coat by a german label ip privacy is not just a fashion statement it's also an invisibility cloak the pattern supposed to confuse facial recognition cameras. when the camera's comes at it recognizes this is a face and this is a face it's also recognizes your insides but that's not too much of a problem it's like telling the algorithm that your human. but you have 10 different faces. it doesn't make sense to the algorithm a tool and so it messes with the learning process is behind the algorithms. georgina rowland's and and a heart from the british art collective dazzle club also want to be undetectable to cameras they paint their faces in different patterns called c.v. dazzle created by artist and researcher adam harvey these shapes are supposed to trick facial recognition algorithms they use smartphones to see if it works i can
10:28 am
see where i'm hidden it's not detecting me say they're the shapes of life. every month the collective organizes a silent march through london the british capital has one of the highest concentrations of surveillance cameras worldwide. the artists paint their faces to protest the lack of transparency behind the collection of biometric data. you make yourself very visible to be invisible so we talk about hiding in plain sight. of his face online allowing anyone to download it his you mean mosques can be used to hide your own identity and confuse the surveillance system what's your opinion on facial recognition would you put your face on the line in return for more comfort join the discussion on you tube or facebook so you assume by.
10:29 am
saying. it is known as the poor man's cocaine tramadol. in ivory coast more and more young people are taking the painkillers. the drug is cheap but often impure and almost always comes with a high price of hard. security. 77 percent. do. not like. the sound of the future electric motorcycles looming. meet tesla. currently live for 602 songs. which even reduces
10:30 am
wins the challenge of. 60 minutes. i'm not laughing at the germans because sometimes i am placed on nothing with them and even germany thinks deep into the german culture. you will take his grandmother day out to you it's all about who you know i'm rachel join me to meet the gemini on the gulf coast. hello and welcome to this edition of the 77 percent i am my committee and it's great to have you with us. coming up on this week's show. we find out why so many young people picked a good drug trial model. to meet a nigerian musician who has become a local celebrity on the italian island of sicily.
10:31 am
right 6 us around.

24 Views

info Stream Only

Uploaded by TV Archive on