EU lawmakers approve world's first legal framework on Artificial Intelligence | DW News

now EU lawmakers have just approved the world's first legal framework on artificial intelligence the ACT aims to balance the safety and reliability of AI use for people across Europe whilst also fostering its development in various Industries it's been five years in the making and it could be seen as a precedent for countries struggling to control ai's rapid advance and for more I'm joined by our correspondent Lucia Shulton who is in Strasburg L Louisa tell us more about what the parliament just voted on yes so the European Parliament today has decided on the EU AI act

and this piece of legislation is a decisive one for regulating the sector of artificial intelligence um it is deemed to be the first one uh by a major regulator worldwide that is that comprehensive and the principle which underlies this um this act is relatively simple so the idea is the riskier the AI system the higher the threshold for regulation um and that goes as far as there are certain things that are being banned by these AI act and then from there they decide there like certain um areas where they decide um what kind of Regulation

is needed and to for this act to enter into force it has been decided today by the European Parliament but it still needs a final approval by the EU member states um which is supposed to happen in April and only then it will enter into force and so give us a sense of what's going to change uh once it does enter into Force what's going to change for people living uh here in the European Union yeah so for people in the European Union the European institutions have been saying they do this they do regulate AI

systems to give people trust into these and there is also a certain amount of clarity coming and I have already mentioned these bans and just to make this clear there are certain things which will be prohibited in the European Union one of these is for example that you will not be allowed to do a facial recognition on CCTV live streams except for the authorities in very very restricted circumstances for example when there is a case of terrorism or if you're looking for a victim of kidnapping but aart from that this will be forbidden as well

as um certain manipulative Behavior there's the the the idea of the social scoring should be banned in the European Union once this enters into force and these bands they are already also expected to enter into Force 6 months after the law has entered into Force so this could be before the end of this year Lucia Shulton reporting from Strasburg thank you so much for that and I'm joined Now by yanos delka who's DW's Chief technology correspondent welcome yanosh um tell me what will this law change for people like you and me living here in the

EU well it won't change anything immediately but we will see increasingly those effects over the next coming years now after this vote today there are still a few formalities left to be settled once that's done that's probably by May or June of this year um the law is going to take effect gradually over the next two to three years and along the lines we're going to increasingly feel the effect of those laws for example for the two of us living here in the European Union we will have a right to know whenever we interact with

an AI system so for example you know if you use a chatbot like jat GPT online there would have to be a disclaimer saying okay this is an AI system you're interacting with the same is true for a call center for example um and then of course there are also changes sort of like less obvious changes behind the scenes that we might not feel immediately but that have you a significant impact on us as well and that is sort of the idea that all the AI systems we're going to use will have to fulfill these

requirements for what the EU now calls trustworthy artificial intelligence okay so quite comprehensive this is we have to remember this is the world's first comprehensive legal framework on AI isn't it so you know how easily and successfully can this be implemented in Practical terms yeah I mean passing laws is one thing and then enforcing them another yes so um you know this is really going to be a litmus test for this the next couple of years um the EU in Brussels is setting up its own AI office that will monitor particularly um you know big

AI systems and then each of the 27 member countries in the EU um is naming their own AI Watchdog where people will be able to file a complaint now you know this is completely new all these offices are being set up now um you know they we will have to see if they get the right people with the right expertise and you know the big question is really you know will they will they get the job done so kind of a work in progress I guess now this is EU wide isn't it does this act

have any influence at all outside of Europe well you know that's at least what lawmakers in Brussels are hoping for so they're hoping for what they call the Brussels effect essentially the idea that you know you the laws that have been passed here will sort of inspire or will be some kind of blueprint for the rest of the world where other you know legislator are now looking into similar issues um there's sort of precedent for that um you know that's what happened with the gdpr our current data protection rules here in the uh in the

European Union they took effect back in 2018 and then we saw that you know afterwards countries like Brazil for example passed data protection laws that very much aligned with the gdpr and well you know now the Hope by many lawmakers in Brussels is that history will sort of repeat itself here okay so leading by example apparently not everyone's happy with some exceptions um that have been made in this law for certain certain applications can you give us a couple examples of those yeah yeah and let me start by saying you know artificial intelligence poses significant

risks to our fundamental rights and the AI act now really addresses a lot of them that being said there are you know know what human rights activists call loopholes in this law um a good example is is the use of um facial recognition technology a specific use live facial recognition technology in public spaces now in general this AI act bans that but there are exemptions for um for law enforcement for example you know when they're looking for a missing child or when there's imminent terrorist threat um now law enforcement says you know we need that

in order to do our job but on the other hand you have human rights activists and um you know people coming from that sort of privacy background who are warning that you know this could sort of erode fundamental rights here and that it would essentially affect everyone living here in the European Union fascinating DW's Chief technology correspondent yanos delka thanks so much for those insights really interesting now artificial intelligence as we know promises many Innovations some of those are feared many are seen as making a positive contribution researchers at Australian University have now developed an

AI application that they hope will benefit people with dementia called Viv a character who serves as a companion for those with memory problems Viv is currently being tested at a care home in Sydney where she Comforts patients who are feeling low now ai companions like Viv are not meant to replace human caregivers but rather provide additional support forgetful and get lost in my thoughts