this plan, the g as a reminder of our top story. ukraine and presidents lot of mr. landscape has concerns that count to offensive operations against russian forces are on the way. it is fighting is being reported in principle. we shall reach and by both keys and most kept coming up next on dw scientists training a i to in type and brain waves the technology. so it'd be able to read our minds. well, find out next on shift. i want to get down for me and the team, thanks for watching. the one of mankind's oldest ambitions could be within reach. what do you see? it really is possible to reverse the researchers

and scientists all over the world. for you know, race against time. they are peers and rivals with one daring goals to help smart nature. the boy likes watching it on youtube. dw documentary come have a page this out. take the highlights you every week in your inbox, subscribe. now when a i read online, soon, scientists have already trained an ai to interpret your brain way. also, experts argue that a poet, beauty fills a potentially dangerous to the point of making you sick. we take a look at snapshot this morphia and how it comes about and why that's natasha. use of data with us intelligence agencies and who else to month use information from

big tech companies. these are the topics that have move the tech world the train day. i can pick your brain and recreate the image as you have looked at all stories that you've got, listen to all by interpreting your brain weight. those are the findings of scientists in japan and the u. s. i in the future fulfill our every wish. as soon as we think it, and how do we get the best results out of ai tools? now, at the university of was stuck in japan. participants in the study looked at thousands of pictures and set the brain scans reported. those guns were used to train in a dakota model, it load, which gun represents what picture they showed. the mother was able to recreate the image is just by studying the brain with of participants why they were looking at images. and it doesn't just work with picture researches in the u. s. trained in

a model using friend scans of people and listening to pop costs and model of been managed to create a somewhat similar text to what they had listened to in process but sofa. it's only works if the model is trained on your individual brain activity. if you cooperate, if you wanted to, and you could step this, think about something else and create different brain with that way. so a one full full a while to streams anytime soon in order to communicate with an a and make it work for us. brain with won't help us right now, but i can already help you a lot. if you feed it with the right prompts, how to prompt, there's no secret magical prompting dictionary assistance, a constantly learning and evolving. yes, how you get better results prompt and english. even though a lot of a, a model speak several languages, they've received

a lot of training and english off the ai what needs to fulfill a task. you can also ask the model to write prompts for another, a i an image, generate to, for example, use chain of thought prompting. give me a, an example of how you want to solve a task at a step by step description of the thoughts you wanted to include this way. you might be able to get better results. but always keep in mind that a lot of models have flaws. which lots to keep in mind. a, i knows only what it's been told. chapter peachy was only trans with the text published before september 2021. so it's contra related to recent events. is biased because it's training data often is prejudiced against women. all minority is like the eligibility q community find the way from the training data into the results. when image generates michonne is asked to create pictures of a doctor. it will be male, acute in person. instead, will be female,

so always check if it results on defensive. it doesn't keep your data safe. samsung employees as cheap each, each fixed improve coat, they uploaded it, and by accident the corporate secrets that could now be included in the checkbooks future responses. the same applies to your personal data. in many cases, your input is stored and used to further train these models. are you using a, a tools in your everyday life already? and what's your experience with prompting flu, this in the queue to know off the lips, face altering apps like face, june of face, of like getting better and better as producing real looking photos and even videos . but some experts argue that a public field a potentially dangerous to the point of making you sick. what's the fuss about face to name today. all you need to enhance just sales. this is an app phase 2 and one

of the most popular beauties field trips is counting more than 200000000 downloads worldwide. and that thousands of others face up. you can make up to 2 plus airbrush the list goes on and on in social. absolutely. take talk or snapshot has been offering integrated beauty filled us for quite some time being used on a massive scale. here's what the study that was conducted in the us 2021 says 80 percent of gross under the age of 13 has already use the filter or re touching apps to change the way they look in the photos. ok. but how does wanting to look pretty and pick just make one sick? snapshot this mafia. the term describes the needs heavily edit one's own digital image, up to the point where you have plastic cetera in real life to match that digital image. the term was coined by purchase position doctors to john asheville. he noticed that an increasing number of cosmetic surgery patients were bringing heavily edited itself is through the consultation appointments. well,

this phenomena is a rather severe case of body to small fia, but it shows what damage this display of supposedly perfect human beauty on social media can do. it leads to completely unrealistic ideas of what you should look like, and that can cause anxiety and depression. cosmetics such as have reports of the patients who bring in heavy ext itself is often surprised that the all the photo graphic results cannot be replicated in real life. because it's so easy in an app. so how do they work anyway? how does face tuning work? the future, so essentially automated photo editing tools to use artificial intelligence and computer vision. the software detects the face and then overlays and invisible facial template. it consists of dozens of thoughts to create a sort of mesh. once that has been built, all kinds of graphics can be attached to the mesh pending on the app. this can be adding some digital makeup reship sets and spatial features or adding some doubles

on what is being done about it. while the haven't been attempts to band is f sofa from european countries are now taking steps to at least regulate the use of a beauty of filters. they are trying to fall. social media advertises and influence so as to admit when they have altered their physical image, especially instances and celebrities are presenting themselves and assuming they authentic wait online and use us often believe that what they see is 100 percent. real. no, i introduced in the in 2021 celebrities. some instances must indicate on social media whether a photograph has been re touched in the process, someone to, to not to go even further and encompass all content on social media. frances, in the process of passing a similar lot, but for both photos and videos. french one edition, we know the mass at the measure is to limit the destructive psychological effect of filters. and the u. k. is currently working on drafting and always comparable tens

. apart from the state regulations, it in fact really comes down to us uses. we need to constantly remind ourselves that what we see online is very often not a depiction of reality. what do you think of these apps? are they a threat? how do you feel about such altered faces? let us know. what are you using in stuff, facebook or what's up? well, be wary of the data. you shift the parent company metric with cost on person. the dates of european uses to us all for a tooth. that's why the social media chime has just been fine in the you with the records, some of the 1.2000000000. why would met to allow sensitive data to be accessed by us government services? what else is this data being shipped? who exactly is being effective? let's take a look how the record find helps you make a transpose that stays health, excuse us, to the u. s. why the company claims slides or to the business model,

use the data needs to be bundled to processes in an effective way. and to said targeted as based on the results for us to valence last, a low intelligence agencies like the end, it's a broad access to data from non u. s. use us a few p union costs auto competitive take precautions against these last before, but it didn't happen. so now that you have left mess up with this record fine in order to protect your pin citizens data from the u. s, it could be an important step and false big tech companies to protect our data better even beyond the you. that's important because it's not just u. s. also, or issues that may be interested in your data. your own government could be to us big tech aiding the n s. s. mess surveillance program. we've known that since 2013, thanks to with the blow. edward snowden. now, 10 years later, government and intelligence agencies are reported the receiving even more data from social media provider. who else has metro. sharing data with mattel has been

publishing and transparency reports in 2016. it detailed the number of times governments of possible access to the own citizens data on its platform. in the latest issue, the us is popping the list with more than 64000 request photos closely by india. and next comes gemini and brazil with about $70500.00 total request submitted each in the report metro state measure response to government requests for data in accordance with a particular law and our terms of service each and every requests were receive is cassidy reviewed for legal sufficiency and we may reject or require greatest specificity or requests that appear overly broad or big. so in other words, unless a country slot does not allow for use of data to be passed on much, i may indeed, and it over the published number seemed to indicate that method is actually quite willing to give out information in more than 75 percent of all cases,

at least some form of data was produced, which states i do governments access beyond basic data like user names, addresses and contact information tech companies like metro as google, apple, microsoft and amazon can often access a lot more. for example, use of emails, text messages, call locks, photos, videos, documents, contact lists, and calendars that think of government or storage is a very intimate look into your private life. what is this data of being requested and used for? in many countries, law enforcement agencies monitor social media to assist with criminal and civil investigations if they have good reason to believe someone's involved in criminal activity. they might ask platform providers for access to that persons use a profile. this can also happen in connection with public safety. for example, before big events authorities might want to scan use us profiles to assess risk. yet another reason is immigration and travel screening. and here we can get an idea

of the ethical implications that come with platforms for meeting use of data to authorities. imagine those requests targeting certain activities. you might think this kind of surveillance doesn't apply to you because you've got nothing to hide. but government monitoring of social media data can and has led to people being accused of crimes even wrong for them. and it's definitely in this free communication. what can you do to protect yourself? the recent loading and the you, the parts of the record fine says match a must return your pin, use the data to your pin service. but apart from that, that's nothing you can do to be entirely safe from being monitored by your own government or elsewhere. you'd have to come see if they did your online accounts to guarantee that. and this does not apply to meta and on, but also to google, amazon and microsoft. what do you think? do you get if your government has access to your data? that's it from me by and see you next time, the

choosing the joy of the right over the annoyance of traffic gunners b like revolution goes to verify the thing is there's about you but. but now that unions was checking out the 77 percent on d w b y for some drivers. what troy cause that is testing out with bmw of u. s. 1000. this to time supervisor to enjoy the

total read. the 60 d w. german was sing for winning offer is available world wide and for every language level, learning german has never been simpler. german to go the welcome to the 77 percent programs. on this show, we talk about the important things for africa. i am michael and i'm excited to bring you this edition from across the street show is inspired by you actually you've been talking to us on social media about it's.