1st time since 3 hostages were mistakenly killed by, as really, soldiers and guns are called the incident. heartbreaking but found to press on with the war against a month and we will end it there. up next, our tech show shift examines with artificial intelligence, can help protect animals, threatened with extinction. that's after the break of michael locus from all of us here. thanks for watching. induced state to and if you can, the rules we say they're about never giving up every weekend on dw, the crew to clean and ha, empire of crime,

news that you can have to ask those one queen because one point a made up currency . we want to be the number one for the car in cl do. then she disappears without the trace on the top 10 list of cartel leaders of financial thriller about the world's most one city woman crypto queen starts to assemble associates on dw. we're living in an area of mass extinction, ones, unmatched. and the last 6 to 6000000 years since the dinosaurs were wiped out over a quarter of all mammals species and one in a broad species on the brink. but can i help save them today on shift? it is not only natural, these losses and climate change. the threat of wildlife poll shows, for example, to over $20000.00 african allison's every year ranges and national parks have been struggling to prevent the practice. but that's non profit organizations,

heck the planet has developed and the camera system to detect poachers and sound the alarm in real time. and a trigger wanting us might find some of the following imagery. distressing african elephant populations have fallen drastically in the last century. one major cause poachers, they want their ivory to stuff them, the nonprofit organization, the heck the planet has found a way to modify the existing camera trips these uh, these cameras, we call them the dump cameras. we. we make them smarter, so we modified, i'm a little so um, so that they are unable to communicate with a box that we created. we manufacture these and make them able to communicate and send the photos wirelessly and analyze these photos using ai. so we have a machine learning algorithm on there analyzes the photo and then we immediately know what's on the photo. if it's a human arts and alison, if

a person is detected and alert a sent arrangers immediately this way, they can react quickly to potential poachers, which wasn't possible before these range are units. oftentimes they just find an edison's carcass. uh and you know, then then the boat years are already gone, like maybe one week or 2 weeks ago. and they don't have any information to go from . because it, you know, they don't know where they came in or where they went out. because it's such a huge area, the, the smart devices make is much more likely to catch intruders in real time. they work via satellite and don't need why fi? the country's like come on. sandy, a, as in bob way had already implemented the smart camera trends. for the endangered african elephants and those who protect them, the technology could mean the difference between life and death. in

india, amazons, and killed a different dangers to around 20 of them. i ship and killed by trains every year. that's why in a i warning system is being installed a long railroad tracks. sense those identify migrations on the ground caused by elephants and sent a warning signals to prevent collisions. it's undoubtedly an issue, but one was by the threat of global warming. 2023 was characterized by devastating fires around the world. greece, spain, hawaii, canada. in australia i, technology is now being used to help protect animals like the kangaroo and the koala. from the effects of bush fires, a koala is on the move under the cover of night. just like these wallabies, the kangaroos, smaller cousins for australian wildlife conservation. these images are a small victory. using camera traps,

they are trying to find out how well different animal species are recovering from bush fires. so the camera takes the animals by a combination of movement and hate from the animals. so as the animal looks past the camera, usually a man with a bite station is if it gets triggered and takes 5 quick photos of the animal as it comes past guess on the air. in greenville researches in the conic blue mountains west of sydney. the area is a tourist drive, but it's also prone to bush fires. that's why it was included in w w. f. us really is large scale camera sensor project eyes on recovery. about 1100 camera traps were installed all over the country, especially in areas damaged by the bush fires of 20192020. the cameras generate millions of photos helps researchers analyze the data. ready

the artificial intelligence is actually enabling them to sort through these images much more, much more rapidly. so it's basically improving their efficiency so they can go out and do you know, the work of the work that they need to do to manage that have spaces recovery, for instance, after collecting the photos from the cameras, researchers upload them to an, a model called wildlife insights the google powered platform has a database of over 35000000 images and is being used by wildlife organizations all over the world. the detect what kind of animal is seen within an image and labels that accordingly helping research move along more quickly. ready a we find ourselves in this point in time where climate change is is meaning we're having more frequent fires, not just in australia, but across the entirety of the world. so projects like eyes on the coverage that

develop technology that can help us to assess impacts more rapidly just absolutely crucial for understanding future fire events in australia and across the world. well, the blue mountains look lushin green. now, it is just a matter of time before the bush fires return and the animals suffer the, the system won't save the animals in the event of this aspect, but knowing where they live and how many they are is extremely important for conservation. but how do the databases work? we asked them to explain it for us. it's as easy as drag and drop. the researchers upload photos to one's life insights within seconds, image recognition, i identifies the animal. so when a research or a biologist uploads their camera trap data to the wildlife insights platform in our

artificial intelligence model looks to be images and based upon what we, what we shown it before and makes, it predicts on what it thinks it sees inside the image. the i matches the photos to animals that has been trying to recognize the database contains entries for over 3000 animal species. a value of one means absolute certainty. so 0.99, just from this image of an elephant, the trunk is very close. indeed. va isn't always that confident. for example, when distinguishing a leopard from a judge, you are on some platforms the i can even recognize individual anonymous sharks, for example, of unique patterns on their skin, which allows them to be identified algorithms, then, much to patterns like facial recognition software. the wireless can be clearly

distinguished by their tail fins, coloring researchers to learn more about their movements, backs and turn can lead to better conservation for the marine jobs. but sometimes pictures alone, the enough waves, for example, rarely come to the surface. that's why scientists also use sounds to find out more about them. sounds like this. what do you hear a, uh, the cause of various ways of specie is recorded and alaska sound samples like these collect and worldwide the problem. microphones, folding victim to interference. but with the use of a i, we so it says that able to filter at the end of here is out. and here they called and our field is mostly an efficiency thing. if you apply it a, i, it kind of figures out the easy decisions for you, and then it lets your expert look at the areas where they eyes and so sure. and it

allows the experts do with the past to validate. and the a, i really takes a lot of them like brute force out of that. so we have estimated that it saves as much as like 96 percent of our time. the sounds can even be assigned to individual way of specie is sped. why is this important? well, some specie is like this right away, and here, critically and agent, we found a call like just kind of accepted call. we didn't know who it belonged to, but we thought it might be a, essentially a right. we'll call wearable the surcharge. hired is we would never been able to do that without a i and for a population of 30. no, every data point counts because they're super rare. you really don't expect to find them. so having, having tools allow you to call these huge datasets we've noticed on terms is essential for populations like that. but it's not just the ocean designs that are threatened with extinction. insights in particular addressed european be

populations are declining sharply, but a i can help you too, by showing us what these really need to try to design their own garden. it would look like this. at least according to alexandra daisy ginsburg, the garden in front of berlin's natural history museum is her art project. as an algorithm, as chosen which plans start to grow here? every time we run the algorithm to create a new prompting scheme, the algorithm is optimizing the taunting choices and arrangements to have the most diverse group upon and 18 species possible. so these were most and many more. following the algorithms, certifications, ginsburg entertain, planted 7000 plants here in the heart of berlin. the project is called pollinators, half maker, to develop the algorithm, ginsburg,

and her team. we searched the exact needs of local pollinators. the algorithm then generated a planting schematic with $150.00 plants species. it even calculates how the garden will change throughout the year. the, the algorithm chooses plots that are appropriate for each location, set the soil conditions, sunshine, so on. and then it arranges, it takes a subset to create unique schemes every time. i'm then arranging them in ways that different charging style up to my businesses of all the different insect pollinators locate their food in different ways. some insects, such as bees memorize the locations of flowers and developed efficient flight routes to visit thousands a day that i will look most kids if you put the algorithm, anticipates the pollinators flight path. so we'll see, that's what the planting pattern is based on the data for the men special alga,

legal voice. oh, it's a, it's a by, it looks unusual to the human eye, but it's designed around the weight and sex. moved around the garden in the garden, just i don't answer. then you can use the algorithm yourself to calculate a plan for an insect friendly garden, free of charge. the web based tool is optimized for your opinion, wildlife, but models for other regions could soon follow hey, i can go through huge amounts of data very quickly helping us understand local greenery and even large and complex eco systems. there's new technology has become an integral part of conservation. that's all for me. take care to next time by the

welcome back to the 7 to 7 percent street debates this week. we are in the car, come refugee, tell us what we have to do is have society rethink how they've handled refugees out there in the web. really skipped because of us that we did say that to begin steam together. honestly, we just want to have a whole the 77 percent of going on d, w. b, y for some drivers. what troy cause that is testing out with bmw s 1002 times super bye. who enjoys the total read

the 8060 minutes on d w. these places in europe, us flashing the rank one step into a bold adventure the treasure map for martin clo, to discover some of us regular bragging sites on youtube, and also the hello and welcome to another edition of the 77 percentage of the show that gives voice to app because you, i'm your host to live so i can't imagine what it must be like to be forced out of your hometown or a country because there's a war going on or because prolong droughts have made it impossible for people to grow crops or feed the animals, so that means that is the harsh reality of millions of people in sub saharan africa . so in this show we'll dive deep into the.