TIK – narzędzie do zmian na lepsze
This is an AI transcription.
00:00:10:05 - 00:00:16:01
Abigail Acton
This is CORDIScovery.
00:00:16:03 - 00:00:39:23
Abigail Acton
Hello. Welcome to this episode of CORDIScovery With me, Abigail Acton. There are so many stories of how damaging our digital lives can be from promoting unrealistic lifestyle goals to the skewing of elections. So this episode, we're shifting the narrative. We'll be taking a look at three applications that make a real difference to people's lives. This is where our impulse to perform random acts of kindness meets the power of AI.
00:00:40:01 - 00:01:02:10
Abigail Acton
All three guests whose projects have been supported by the EU's Horizon 2020 program are here to share the results. A very warm welcome to Christian Erfurt who launched his first healthcare startup while still in high school. Christian is co-founder and CEO of Be My Eyes, an app that pairs sighted volunteers with visually impaired users to help them deal with day to day challenges.
00:01:02:16 - 00:01:03:22
Abigail Acton
Good morning, Christian.
00:01:03:24 - 00:01:06:01
Christian Erfurt
Good morning, Abigail. Thank you for having me.
00:01:06:03 - 00:01:20:16
Abigail Acton
Christos Ntanos is research director at the Decision Support Systems Laboratory at the National Technical University in Athens. He is the coordinator of the ChildRescue Project, a collective awareness platform used to find and rescue missing children. Welcome, Christos.
00:01:20:18 - 00:01:22:06
Christos Ntanos
Hello. Thank you for having us.
00:01:22:12 - 00:01:40:11
Abigail Acton
Francesco Saverio Nucci was the coordinator of the FANDANGO project and Italian mathematician and researcher and innovation manager. Francesca graduated in Artificial intelligence at the University of Rome. La Sapienza. Here's Applications Research Director at the Italian company Engineering. Welcome, Francesco.
00:01:40:13 - 00:01:42:16
Francesco Saverio Nucci
Thank you. Nice to be here.
00:01:42:18 - 00:01:56:12
Abigail Acton
Excellent. Well, welcome to the three of you. And I'm going to turn to Christian first with my questions. Christian, BEMYEYES pairs visually impaired users with volunteers who help them with specific tasks. Could you tell us how it works and how many people are using it, actually?
00:01:56:14 - 00:02:17:08
Christian Erfurt
Absolutely. BE My Eyes is an app connecting blind to his computer people to sighted volunteers by the press of a button to download the same app, sign up as a blind user and we match to a volunteer in an overlapping timezone in the same native language as you speak. So you don't have to worry about disturbing someone. And we call in batches of 8 to 10 people.
00:02:17:08 - 00:02:41:07
Christian Erfurt
So the first one available is actually someone who actively decided to step in and take the call. It's a two way audio. One way video. Backside of the camera is turned on and you can see with the blind, obviously people a person are struggling to see. It's currently in 150 countries. We have five and a half million volunteers who have joined the platform over the years and 350,000 minorities compared.
00:02:41:07 - 00:02:44:14
Christian Erfurt
People have signed up since we launched in 2015.
00:02:44:16 - 00:02:54:23
Abigail Acton
That's an incredible number. It's very, very impressive, particularly the volunteer base. That's huge. What do you think the motivation is for volunteers? Why are people signing up?
00:02:55:00 - 00:03:16:02
Christian Erfurt
And I think that there is a very clear call to action for both sides of the equation. And you can clearly see it from the blind person's perspective. But for the volunteers, it seems like it's being part of the solution. You know, to do that small micro volunteering act of kindness and it's very gratifying to be part of that exchange.
00:03:16:04 - 00:03:20:17
Christian Erfurt
One of our users actually said it feels like a good deed waiting to happen in my pocket.
00:03:20:19 - 00:03:30:02
Abigail Acton
Excellent. Yeah, that's a good description. I can imagine that. And also, I suppose it's really seamless. So the phone goes. If you're the volunteer, you decide whether to take the call or not. I guess that's how it works.
00:03:30:02 - 00:03:52:18
Christian Erfurt
Yes, exactly. It will happen and it will ring at some point. And if you're in a place and time for for answering it, you can can answer and help with anything from distinguishing products, shopping, navigation, you know. Technology is great, but it can only take us so far. Google Maps will take you to a location, but not necessarily show you where the door is.
00:03:52:20 - 00:03:56:16
Christian Erfurt
And so we've seen all kinds of use cases all around the world.
00:03:56:18 - 00:04:13:08
Abigail Acton
Right? Absolutely. And so the blind person uses their camera to show what it is that they're struggling with and gets talked through the solution, etc.. That's excellent. So, Christian, is it free for the blind users? I'm sure the volunteers don't have to pay anything. How does this work in terms of monetization or is there no money involved at all?
00:04:13:12 - 00:04:42:16
Christian Erfurt
It's a very good question .BEMYEYES is absolutely free for all of our users to blind and visually impaired volunteers. We've always set out to to have a product that would be globally available and free for all of our users, which is a challenge to we have to keep the lights on too. How do you come up with the business model and we have a very strong design center saying everybody wins only when nobody loses, and we dared to aim to build a business around that design tendency.
00:04:42:18 - 00:05:11:22
Christian Erfurt
And for the first three years, we didn't know how to do so. But actually it was a collaboration with Microsoft that led to the business model and that is that today be my eyes. This is enabling companies to provide accessible customer support and the companies are paying us to do this to to do so. So they simply gain access to our community and we enable them to have a video line established directly into their customer support center via our app.
00:05:12:00 - 00:05:35:01
Christian Erfurt
So you could say that it's accessible customer support and we're providing the video connection for them, which makes it a lot easier to troubleshoot a product. It could be. You know, it's difficult enough to install a router for a sighted, but but if if you know what the latest blinking is is the power on all of that the troubleshooting is just a lot faster when they have the video model in place.
00:05:35:02 - 00:05:58:01
Christian Erfurt
So Microsoft was the first customer to join for all of their Windows products, all of their and Microsoft, any kind of product, really. And I think what was fascinating was that immediately they invited Google to come and join and see how they were doing it. So they didn't want to use this as a competitive advantage. They said this is such a great movement that we need more companies to join.
00:05:58:02 - 00:06:03:10
Christian Erfurt
So they invited a competitor into, you know, the belly of the beast of their customer support center.
00:06:03:12 - 00:06:07:00
Abigail Acton
Wow. That's really impressive. You're right, because that sort of stuff is normally jealously guarded,
00:06:07:02 - 00:06:30:18
Christian Erfurt
Exactly. And these have been some of our biggest advocates in getting new customers on board. So today we're doing it all in. We're doing health care and transportation. We're doing a tech support e-commerce. And I think it's because of the companies are not only solving the problem and it's faster than the traditional way of just an audio connection.
00:06:30:20 - 00:06:43:08
Christian Erfurt
But more interestingly, they get an insight to whether the products are not working from an accessible perspective. Right. So you could say it's an insight to the customer journey and what they need to fix for the millions who are not calling it.
00:06:43:09 - 00:06:46:14
Abigail Acton
Right. Right. So that's great because it's like real time feedback,
00:06:46:16 - 00:06:47:09
Christian Erfurt
Exactly.
00:06:47:11 - 00:06:53:01
Abigail Acton
Yeah. That's fascinating. Thank you. Has the success of this concept inspired you to think of developing other tools?
00:06:53:03 - 00:07:24:16
Christian Erfurt
I think in general that the combination of purpose and profit is absolutely probably the strongest competitive advantage going forward. All of us as consumers, we want to be part of of of solutions rather than problems. So I think that this should inspire others to do purpose, drive it as a combination in a state of be my eyes. We are sliding into verticals such as colorblindness and dyslexia and also aging eyes supporting people who just don't necessarily understand the situation they're struggling with to see.
00:07:24:17 - 00:07:30:15
Christian Erfurt
So but I see I think we will see a lot more of this micro volunteering movement in the future.
00:07:30:15 - 00:07:46:00
Abigail Acton
Well, I think you're right. It taps into something quite fundamental within us, I think. And it brings the power of of digital communication to to that, making it so much more so much more widespread, so much easier to do a lot of good rather than just a a small saying next door for a neighbor or something. Fantastic. Excellent.
00:07:46:05 - 00:07:50:01
Abigail Acton
Does anyone have any questions for Kristen, please? Yes. Christos.
00:07:50:03 - 00:08:10:14
Christos Ntanos
What question? I'm really fascinated by your project and the the results it has. And I really wonder how do you validate the the volunteers in order to be certain that they're safe to provide their services to the the ones that need it?
00:08:10:16 - 00:08:37:23
Christian Erfurt
It's a great question and thank you for asking it. We actually give everybody benefit of the doubt, so we don't validate we don't do a background check for them to do. And anyone can have a chance at joining the human rights community and we will, you know, act on any abuse reports. But this is a beautiful engagement that happens and it happens, as I said, in 150 countries and 185 different languages.
00:08:38:00 - 00:09:04:07
Christian Erfurt
And and we rarely see any kind of problems with the the input from the volunteers. You know, we do see some prank calls from one time to them that someone will sign up. And but I think the community is really defending the model and saying you're doing it for the money instead of blaming FEMA. There's a concept which is I think is also a powerful thing to have, is that your community is kind of defending the very basics of how it works.
00:09:04:11 - 00:09:10:20
Christian Erfurt
So we don't validate it. We don't background check anyone. They get the benefit of the doubt. Everybody is welcome to join.
00:09:10:22 - 00:09:24:12
Abigail Acton
Well, given that it's over the telephone as well and it's not immediate contact, so prank calls, I imagine in U.S. schools is pretty much where, you know, as bad as it could get. No, that makes a lot of sense. I like the idea of peer to peer regulation as well.
00:09:24:17 - 00:09:51:14
Christian Erfurt
And you can say that because you don't have you can introduce yourself with your location and name and everything. But but it is actually, in some sense anonymous because you connect and someone who speaks the same native language as you do will answer the call, but they don't have to know anything about you. You can just ask for, you know, product distinguishing or I'm saying I need to call a coordinate this outfit or find the entrance to the metro and then hang up.
00:09:51:16 - 00:10:11:13
Christian Erfurt
But we also see the communication we're coming back to and the the greater good of it is that they understand what it's like to have a visual impairment as a volunteer. And that might make you more and you might approach someone in real life more easily. Now that you know that you know you're not that different.
00:10:11:16 - 00:10:31:07
Abigail Acton
Yes, indeed. Bringing down barriers generally. Excellent. Christos, I'll come to you now. Thank you for that last question that you asked. And I know you're in the chair. The Child Rescue Project created a web platform and a mobile app that uses machine learning to help organize agents locate missing children. So, Christos, there are alert systems in place already.
00:10:31:08 - 00:10:33:17
Abigail Acton
How does child rescue improve on them?
00:10:33:19 - 00:11:07:24
Christos Ntanos
Well, currently, the most widespread system is the Amber Alert system, a very famous one, and it's a very successful system currently in place in most EU countries. The public is notified through a nationwide TV and radio broadcast, but also posters, fliers, electronic message board signs, public announcements and other means. So if a child has gone missing in the public path, apart from the active search by the police and of authorized volunteers, you expect that the information will come from people who have seen a picture in an alert or and have identified the the missing child.
00:11:08:01 - 00:11:33:02
Christos Ntanos
But the alert will play on your TV at home where you may or may not have paid attention to it before or after you go to that park where the child has gone missing. But it's not when you're there. And then we also quickly forget faces in just a few minutes. Most of us can't recognize a person whose picture we just saw, the radio broadcast and the public announcers don't have a picture and neither does the message boards.
00:11:33:03 - 00:11:58:22
Christos Ntanos
And on top of that, you may be watching a picture of a child who has gone missing half an hour ago in a park which is 500 kilometers away. So you should be able to notify the people who are likely somewhere near the child at the time when their eyes those senses are most useful for the investigation. We need to tell people to actively keep an eye out for a child whose life may be at risk right then and there.
00:11:58:24 - 00:12:11:17
Christos Ntanos
And if new information becomes available to be able to tell people elsewhere right where and when it matters most. So this is more or less the the main idea of where a child rescue starts from.
00:12:11:19 - 00:12:20:11
Abigail Acton
So it's like targeting the alert to the people who are most likely to be able to do something about it and targeting it at the right time as well.
00:12:20:13 - 00:12:44:20
Christos Ntanos
Correct. So the main public facing tool of Child Rescue is a mobile app which everyone in Greece and Belgium can get from both Android and iOS. The app simply sends a notification with a picture and a few details. Only if a child has gone missing and is somewhere probably somewhere near the the location of the of the person receiving it.
00:12:44:22 - 00:13:24:13
Christos Ntanos
So this is done without even knowing where your mobile phone is. We send a notification to all the phones and the area we want. The notification is only selected from the phone itself. So that's a technique called geofencing. So we don't need to know people's actual location. So behind the scenes, the child's local platform allows organization for the Missing Children Europe Network to combine information from multiple sources about a missing child and using advanced analytics and machine learning to suggest points of interest for the investigation and send notifications to citizens will have the app around these areas.
00:13:24:18 - 00:13:33:04
Abigail Acton
So can I ask you a little bit what you mean about advance planning? I mean, how exactly does that work? Could you explain that a little bit? Break that down a bit. How does that function? The computer learning.
00:13:33:06 - 00:13:56:21
Christos Ntanos
The way that the application is working is a two way street. So you get information from the the citizens and from other sources, either the police or the usually the family of the of the missing child. And you use that information in order to add to point certain areas on a map and possible routes that may connect those areas on the map.
00:13:56:22 - 00:14:22:11
Christos Ntanos
So you use that alongside other geographical information that has to do with speed of travel based on the means of transportation in that area. And so you either widen the area, create new points of interest and basically pinpoint the locations where possible searches or notifications could be sent. The way that this is a is implemented is always a live thing.
00:14:22:11 - 00:14:34:18
Christos Ntanos
The decision to a human that will make it and using that decision to either send notifications to the suggested areas or send volunteers in that in those areas.
00:14:34:20 - 00:14:54:10
Abigail Acton
Okay, that sounds excellent. I'm seeing something I mean, to to to have an example. Imagine, then, that a child goes missing somewhere relatively close to a motorway. Then I guess the pinpoint would be wider because the speed of distance would be easier because everything's traveling faster, maybe places of interest would be things like service stations and motorway cafes.
00:14:54:10 - 00:14:56:01
Abigail Acton
Is that the kind of thing that we're thinking here?
00:14:56:06 - 00:15:23:03
Christos Ntanos
Yes, it's exactly that. And more even more than that. It has also to do with information that is input into the system, like the common places where the child has been spotted, that have places of interest, like friends, houses, families, houses and points of interest on that route like bus stations and train stations and the like.
00:15:23:05 - 00:15:47:04
Abigail Acton
Okay, that sounds fantastic. It sounds so much more efficient and more powerful than the existing method that you described at the beginning of the Amber Alert concept. So who can so anyone can download the app? As you've just mentioned, it's available for Android and for for iOS who uploads the information. So you're saying, for example, that it's the police and the missing children's organizations that that upload the information to find the child, I guess?
00:15:47:07 - 00:15:49:24
Abigail Acton
Or can the parents of the missing child directly upload information?
00:15:50:04 - 00:16:15:09
Christos Ntanos
We do not ask that kind of sort of treatment from the parents. The coordination of the search is between the police and the family and generally authorities. Authorities are not directly linked to bear to child rescue. So we only ask for citizens awareness from from the mobile app on behalf of the missing children organizations for the country and nothing more.
00:16:15:11 - 00:16:46:18
Christos Ntanos
Each organization is responsible in its country to handle these Internet handles. These these data. The citizens aren't even told what area that notification was sent to. That was to avoid what we call mob justice, or the possibility that you have a group of people joining and then going on and trying to locate the child. And if that's a criminal case, it may lead to certain difficulties that have to do with people taking the law into their own hands.
00:16:46:20 - 00:16:47:11
Abigail Acton
Right. Yes.
00:16:47:13 - 00:17:12:07
Christos Ntanos
So in order to avoid that, I would not give but the minimum of information to that to the citizens. So what we want them is to send facts, as we call them, in the system. And it's any kind of information that is relevant to the case which may be either anonymously or eponymously a photograph, a piece of text, a sighting, a description, anything.
00:17:12:08 - 00:17:42:04
Christos Ntanos
So currently we also use the the platform to coordinate volunteers in Belgium. For them, it's a different app basically that that's a different app specifically for the volunteers that helps them to define active search areas and those in those areas. The volunteers can coordinate searches, including with specially trained dogs that are distributing fliers and putting up posters, pinpointing the locations to retrieve them when the investigation is over.
00:17:42:06 - 00:18:22:16
Christos Ntanos
Privacy is a very important issue for the for for child rescue, because we're talking about children's information. And if it goes out, especially in the Internet, right, then it's never leaves it. So what we do is to add to try to have that information located only in mobile phones where it matters. And when the investigation is over, that information is securely deleted from each and every place, including the fliers where the volunteers know where they've put them up, and they went back there to take them down.
00:18:22:20 - 00:18:38:12
Abigail Acton
No, that sounds like it's it's covering all all possible aspects of of intrusion very, very well indeed, given the fact that it's so robust and so on. Do you can you have some statistics of users, Do you know how many people have downloaded and and how many children have actually been found with this?
00:18:38:14 - 00:19:13:08
Christos Ntanos
Now, we have over 20,000 users in Belgium and Greece, which is very good for a research project, and it's actively used in both countries alongside the standard practices employed by the missing children organizations to assist police. And they investigation in the first six months of the operation of child Rescue between September 2020 and March 2021. We already had it in 40 real life cases in Belgium by the Organization of Child Focus that handles Amber Alerts, though.
00:19:13:10 - 00:19:44:10
Christos Ntanos
And between September and March 2021, that was also for 30 real life cases for the smile of the child in in Greece. Now from the the children that were lost in Greece, every single one of them was found. And from the 40 lreal life cases in Belgium, 32 were found during that time. We don't have all the information because that's again, an issue of privacy and active investigations.
00:19:44:12 - 00:19:56:09
Christos Ntanos
But it actually got at quite a few facts that were sent by the general public, especially in Belgium, where it received 62 facts by the public at that time.
00:19:56:11 - 00:20:04:20
Abigail Acton
That's that would have been of great use in the hunt for those children. It must be immensely satisfying to know that you're producing something with such practical, tangible benefit.
00:20:04:23 - 00:20:23:12
Christos Ntanos
We really hoping that in some way we are managing to add to make the investigation just a little bit faster, because we all know that in the investigation of missing persons, the first 24 hours are the most crucial ones. And so every minute counts when looking for a missing child.
00:20:23:15 - 00:20:29:18
Abigail Acton
Yes, indeed. And Christos you've mentioned Belgium and you've mentioned Greece, Belgium and Greece. What about other countries?
00:20:29:22 - 00:21:04:14
Christos Ntanos
It is currently open to any other country that wishes to to engage with the existing platform and they are free to join. We have started communicating our results to the other members of the missing children. Europe Network and every organization in that network that has the authority to join the The investigations through the child rescue is free to ask for more information in order to get a demo account to start from to see how it works, how it operates.
00:21:04:20 - 00:21:19:02
Christos Ntanos
It's based in a in a more or less nonprofit fashion because it needs to be sustained in some fashion. Since the project is over and the European funding is over.
00:21:19:02 - 00:21:19:17
Abigail Acton
Yes, of course.
00:21:19:17 - 00:21:29:21
Christos Ntanos
So in order to sustain it, there is a minimum requirement that will be communicated when you join the the network.
00:21:29:23 - 00:21:39:06
Abigail Acton
Right? Yes, that makes sense. As as Cristian said earlier, you have to keep the lights on. Exactly. Yeah, exactly. Does anyone have any questions for Christos? Yes. Francesco.
00:21:39:08 - 00:22:19:00
Francesco Saverio Nucci
Yeah.Christos Very interesting. My question is about the people to the action using artificial intelligence and this be sensitive. What is your feeling in the investigation and the and specifically I've seen in my project that that we would talk later about that that there is a in my in my experience a lot. I mean we need is flexibility we can use the machine learning algorithm might give you some suggestion in my experience is the people would like to know why.
00:22:19:02 - 00:22:23:05
Francesco Saverio Nucci
What is your experience of this?
00:22:23:07 - 00:22:53:17
Christos Ntanos
You're absolutely right about the explainability It needs aid needs to to make clear why it asks and why it does something as a way that we've treated it since it's a very, very sensitive issue, is that we keep the A.I. as back in the process and with as many filters by humans as possible. So as we do, say in decision support systems, it's the decision maker who makes the decision.
00:22:53:17 - 00:23:25:04
Christos Ntanos
It's a person, it's a it's an actual person who who makes that decision and not the system itself. We have everything documented as to how A.I. is used in the platform. It's used in a way which only suggests areas of possibly interest, and those areas are then reevaluated by the the police and the mission to the organization. All decisions that are made through the through the platform are made by humans.
00:23:25:06 - 00:23:25:23
Christian Erfurt
I have a question.
00:23:25:23 - 00:23:27:19
Abigail Acton
Yes. Christian, what would you like to ask Christos?
00:23:27:24 - 00:23:31:08
Christian Erfurt
So are all these cases kidnaping?
00:23:31:10 - 00:23:54:09
Christos Ntanos
Well, almost none of them are kidnaping. Most of them are a teens running away from home. And there are certain other cases that have to do with younger children that have gone missing in public parks or when visiting friends or anything, that the the truth of the matter is that their parents should not really be afraid that much.
00:23:54:11 - 00:24:05:18
Christos Ntanos
Of course, they should always be careful for the for the children. But it's true that it's very rare that a criminal abduction takes place with children in Europe.
00:24:05:24 - 00:24:22:11
Abigail Acton
Christos thank you very much. It's it's a wonderful application and I really hope other countries take you up on this, maybe with a little bit of publicity from this podcast. Who knows? Fingers crossed for you and for the children, because it sounds like it's something that that is really potentially a lifesaver. Francesco moving to you now on your project.
00:24:22:11 - 00:24:37:06
Abigail Acton
Fandango has some solutions when it comes to how we can assess what online news to trust. So can you tell us a little bit more about the context? What motivated you to develop a system that can identify suspect news?
00:24:37:08 - 00:25:10:14
Francesco Saverio Nucci
The motivation was the importance to have the right and efficient way to communicate news in never a democracy process is very important to have the right information to decide that. And when we started this problem was not so urgent as now because we started more than five years ago and that we made we see anticipate something that is now very, very advanced.
00:25:10:16 - 00:25:37:09
Francesco Saverio Nucci
And in the middle of the COVID, the pandemic happens. And now we know that this is very important, not just for the for the have, but and it is incredible. So we are let me see the motivation was was when we started. But now is it. Yeah. Even higher. Yeah. So it's very important. And let me say that we absolutely not solved the problem.
00:25:37:09 - 00:25:40:19
Francesco Saverio Nucci
Just started to find some solutions.
00:25:40:21 - 00:25:54:16
Abigail Acton
I like this idea that you realized that this was something that really could go badly wrong in the very near future given the power of the Internet and communications there. Indeed. Okay, so what can Fandango do and how is it used, please?
00:25:54:18 - 00:26:30:23
Francesco Saverio Nucci
Yeah. First of all, it is absolutely in line with the Christos approach because we cannot solve the problem. We cannot change the process within the Germany study. And so we can just support the people that are checking the news and we can just maybe try to give an easier way and a faster way too, because as people say that the them in the in the child the first day, the first 24 hours are important in the news.
00:26:31:00 - 00:26:52:20
Francesco Saverio Nucci
The first 10 minutes are very important because in other case, the fake news propagate so that information can propagate very, very fast. And so is important to to support and a journalist in that facilitate and in a more fast way they walk. But today journalists should be out there in the end. And.
00:26:52:22 - 00:27:06:13
Abigail Acton
Yes, yes, absolutely. It's it's it's the person that makes the final decision. But but how does this actually work? So I'm a journalist. I'm curious to see where the latest news about vaccines is reliable and robust or not. Tell me, how would I use your tool?
00:27:06:15 - 00:27:42:08
Francesco Saverio Nucci
Yeah, we try to define this for different dimensions in the fake news. The first one is the text or indeed in the news and the communication. The first one is the text. So and so the text that we try to use on that should landing in that one language. So analysis of the text to understand if there is something suspect, for instance because you have in this information you such specific features in the text for instance a lot of objectives or something like that.
00:27:42:10 - 00:28:20:00
Francesco Saverio Nucci
Then in another dimension, it's naturally the image. And so we try to use machine learning algorithms to understand if an email address has been manipulated. And during the project, again, that we have seen, that is a deep fake phenomenon. So we work in parallel for image and video to the face, the deep fake, and then but then we have another important part of that is the propagation, the sources and the propagation in the night to create the social networks of the disinformation.
00:28:20:00 - 00:28:47:23
Francesco Saverio Nucci
And then we have seen that there is different kind of propagation so that following fake news, comparing with the normal news. And so we try to understand, to use the emotion and understand how this propagate and to identify some pattern of propagation of disinformation. All this and it this can be used by a journalist when he has some a suspect on need to check some news.
00:28:47:23 - 00:29:07:03
Abigail Acton
So it's like a very precise cross referencing to weed out the stuff that seems to be suspect. Yes, I see that. So but what what absolutist the journalists do there on the computer, what is this, a browser plug in? Is this an application? What do they put the text or the web page into some system that your project then analyzes?
00:29:07:03 - 00:29:08:10
Abigail Acton
How does it actually work?
00:29:08:10 - 00:29:42:17
Francesco Saverio Nucci
Yeah. Yeah, it does. This establishes that website of this, an application under that, but it is depending also from the journalist because when we say journalist is a very large if a kind of of of people that and let me mention that in the project we have three different use so you're starting to see and important to outside contact in Italian newsagency so and this is a and they have some kind of way to walk then we have a broadcaster from Belgium.
00:29:42:19 - 00:30:14:05
Francesco Saverio Nucci
Is this a media professional about working in the broadcaster and they try to include that this business inside the newsroom for this broadcaster. And last but not least, we have a fact checker, a sort of function that is a nonprofit organization in Spain and there is a lot of effort in the also by the European Commission to support a network of fact checker in Europe.
00:30:14:07 - 00:30:19:02
Francesco Saverio Nucci
And it's very important that. But they have a different model compared with the journalism.
00:30:19:03 - 00:30:40:08
Abigail Acton
Okay. So the fact checkers, a major broadcaster and we're talking here not about individual journalists who might be freelancing or autonomous in some way, You're talking about large news generating organizations. Do you have plans to get it out onto onto the desktops of of individuals who might simply be freelance journalists or writers?
00:30:40:14 - 00:31:21:24
Francesco Saverio Nucci
Yeah, we cater to let me say that at the start of the project, we have no idea of of the project that we know what we would like to investigate, but we don't know about the model fully since we don't know if it can be possible to do something that's a business to consumer for everyone. But we did say that this approach to investigate mild to large organization, not to individual journalists I think that this more complicated for individual journalists are even more complicated for control them because naturally everyone say, okay, this is a system against fake news or to do fact check.
00:31:21:24 - 00:31:39:14
Francesco Saverio Nucci
And so I can use that because naturally did it as a job and not just for the journalist, but for entity kind of citizen. But we decided to be more focused on the journalists and on the lines of the organization, even if the facts are not, but is dedicated.
00:31:39:16 - 00:31:52:14
Abigail Acton
Yeah, but kind of makes sense because if you're clarifying the validity of the news upstream, then the news that these larger organizations are disseminating is more likely to be robust and reliable. So that does sort of make sense. I can see what you're saying that.
00:31:52:15 - 00:32:14:00
Francesco Saverio Nucci
Yeah, but even because this is all a symptom of the budget, because we cannot do all but we know that there are more objects in Europe than we've become because in the actual program we have a lot of attention about this. And so maybe put all the speeches together. We're going to do something for consumers, for the citizen.
00:32:14:02 - 00:32:21:04
Abigail Acton
Right. Excellent. Thank you very much. That sounds fascinating. I can see that there is a question from Christos. Christos. Would you like to ask Francesco a question?
00:32:21:06 - 00:32:51:13
Christos Ntanos
Francesco, I'm really intrigued in the search for fake news and disinformation and I'm following up a lot on what's going on with COVID disinformation and organized campaigns from state actors regarding disinformation. And I'm really wondering how do you differentiate between disinformation that comes from individuals and disinformation and misinformation that comes from state actors and organized campaigns?
00:32:51:15 - 00:33:25:10
Francesco Saverio Nucci
Good question. Good question. Naturally I would like to say something about this. And first of all, that we try in the project to be focused on some specific domain, that it was climate change, The European policies and immigration. And this is something that probably has since we're going to receive more from individuals for their for the mediation, that's certainly something that climate change can be understood them from the larger issues.
00:33:25:12 - 00:33:56:23
Francesco Saverio Nucci
And disinformation is a very large problem with a lot of implication in terms of policies. But should that move the relationship between countries, as you know, that's now developed, this is something that we can see when we discuss about that, when we investigate about violation of fake news. So in that sense, so you can have that. You can try to understand if it's something that is coming from different organization.
00:33:57:00 - 00:34:16:10
Francesco Saverio Nucci
But the problem with the fake news is the fact that they cooperate and maybe they going say it's fake news that came from account this, but then it was before lunch and again from an individual. And there is also fake news that came I mean after two years of the same.
00:34:16:13 - 00:34:21:23
Abigail Acton
What do you mean by that Francesco fake news that that came after two years The same. You mean that sometimes it rolls around again?
00:34:22:01 - 00:34:54:12
Francesco Saverio Nucci
Yeah. Yeah. Maybe it's after two years. Or you can have a fake major dove that resurfaces. The center of it is a plane problem in the plane and then you discovered that this was an image of ten years ago, right? Something like that. And then this organization that investigated about this, I'm not so expert about that, but the truth is such newsguard as a sort of archive of fake news with the finger prints.
00:34:54:14 - 00:34:56:16
Francesco Saverio Nucci
And this is can be also being supported.
00:34:56:19 - 00:35:14:02
Abigail Acton
Okay, that sounds particularly useful in this day and age and likely to get more and more useful as time goes on. I think this problem is only going to to get worse. It sounds a little bit like it's a a fight between the AI tools used to identify and the fight between the AI tools used to generate tracks.
00:35:14:05 - 00:35:15:00
Francesco Saverio Nucci
Exactly.
00:35:15:05 - 00:35:32:19
Abigail Acton
Yeah, exactly. Yeah. Well, listen, thank you very, very much, all three of you, for this fascinating discussion. I've really enjoyed this and it's really positive. I think, to focus just every now and then on the use that we can put all of this technology to rather than how detrimental it is all the time seems to be what we hear mostly.
00:35:32:19 - 00:35:37:11
Abigail Acton
So thank you very much for shining a more positive aspect onto this. Thank you to my three guests.
00:35:37:11 - 00:35:37:21
Christos Ntanos
Thank you.
00:35:37:21 - 00:35:41:10
Francesco Saverio Nucci
Thank you. Thank you very much.
00:35:41:12 - 00:36:04:24
Abigail Acton
Are you interested in what other EU funded projects are doing in the realm of ICT digital technology, and how these tools can make our lives better? The Cordis website will give you an insight into the results of projects funded by the Horizon 2020 program that are working in this area. The website has articles and interviews that explore the results of research being conducted in a very broad range of domains from robotics to recycling.
00:36:04:24 - 00:36:22:03
Abigail Acton
There's something there for you. Maybe you're involved in a project or would like to apply for funding. Take a look at what others are doing in your domain. So come and check out the research that's revealing what makes our world tick. We're always happy to hear from you. Drop us a line editorial at Cordis dot Europa dot EU.
00:36:22:05 - 00:36:29:06
Abigail Acton
Until next time.
Technologie informacyjne i komunikacyjne pozwoliły nam na utrzymywanie kontaktów międzyludzkich w trakcie pandemii. W tym odcinku zapoznamy się z trzema projektami, które przenoszą tę zaletę TIK na wyższy poziom dzięki połączeniu potencjału technologii w zakresie łatwego dostępu i globalnego uczestnictwa z charakterystycznym dla człowieka pragnieniem przynależności do grupy i pomagania innym. Uczestnicy tych projektów starają się przyczynić do poprawy jakości naszego życia w różnych obszarach – pomagając w odnajdowaniu zaginionych dzieci, umożliwiając oddzielanie wiarygodnych wiadomości od fałszywych informacji oraz wspierając osoby niedowidzące w stawianiu czoła codziennym wyzwaniom. Christian Erfurt, który założył pierwszy startup związany z ochroną zdrowia jeszcze jako licealista, jest jednym z twórców oraz dyrektorem generalnym firmy Be My Eyes. W ramach wspartego ze środków unijnych projektu BEMYEYES powstała aplikacja, które umożliwia wolontariuszom nawiązywanie kontaktu z niedowidzącymi użytkownikami, pozwalając im rozwiązywać problemy w czasie rzeczywistym. Każdy z nas zetknął się z ogłoszeniami o zaginionych dzieciach. Ale w jaki sposób można by najefektywniej zamienić te papierowe zawiadomienia w narzędzia cyfrowe? Christos Ntanos, dyrektor ds. badań w Laboratorium Systemów Wspomagania Decyzji Narodowego Uniwersytetu Technicznego w Atenach, opowiada o odnajdowaniu i ratowaniu zaginionych dzieci przy pomocy platformy informacyjnej opracowanej w projekcie ChildRescue. Fałszywe informacje wpływają na nas wszystkich i rozprzestrzeniają się w zastraszającym tempie. W odpowiedzi na te zagrożenia uczestnicy projektu FANDANGO, na którego czele stoi Francesco Saverio Nucci, postanowili wykorzystać sztuczną inteligencję, aby pomóc agencjom informacyjnym w rozpoznawaniu podejrzanych wiadomości.
Czekamy na Wasze opinie!
Jeśli chcesz podzielić się z nami swoją opinią na temat naszych podcastów, napisz do nas! Wszelkie komentarze, pytania lub sugestie prosimy przesyłać na adres editorial@cordis.europa.eu.
Słowa kluczowe
CORDIScovery, CORDIS, TIK, BEMYEYES, ChildRescue, FANDANGO, dzieci, osoby niedowidzące, wolontariusze, fałszywe informacje