Introduction

So basically Porto Digital is a municipal company that works with the city of Porto in terms of trying to help with digital transition. We like to call ourselves a toolbox for innovation in digital transition. So we assist the city hall in looking at innovation and technologies out there that can help the city become more efficient, better for citizens, improving quality of life.

Porto's Approach to AI

So today we're going to talk about what is the Porto approach for AI. We all expect, you know, these nice things as we have traffic prediction and flood prediction. We do this, so we do have data we collect from the city.

As you might have noticed, may have not noticed, we can get data from traffic controllers in the city. We get data from the e-scooters to monitor where they work, where they stop, if they are having issues. We have data from public transport, data from energy consumptions, other types of data, Wi-Fi. And we do a lot of things with this.

We also look at air quality as a use case. We try to monitor air quality in the city, try to provide data for universities and the city hall to do some analysis on it.

Use Cases and Applications

In mobility, we're looking at ways to visualize traffic in Porto, as you might have noticed, is difficult. But these are all things we do with analytics.

We also work with AI in terms of traffic prediction, flow prediction. We're trying things to see if we can improve.

Role of PortoDigital

It is not our space as PortoDigital or the city to take the space from the private sector. That's also for them to give us the tools and then we're going to employ.

We are not here for that, we're here to understand it because as you might have imagined, the way these technologies are going compared to the way our city can catch up is difficult. So technology is going way too fast for what we and our technicians in the city can keep up.

Current and Future Focus Areas

So that's why we have PortoDigital here to try to mitigate a bit the gap And what we're currently looking at this year and the next year is the way we curate the physical space. So the cleanliness of the streets, the protection, the safety of our citizens.

We're looking at doing that in the digital space. And that's what we believe is also the role of a city in this area.

Key Areas of Concern

So we're looking at privacy and digital rights. We're looking at bias and discrimination in these algorithms. We're looking at digital inclusion and social justice and digital transparency.

I'm going to try to talk very shortly four topics of this on how we're trying to tackle this and things in concrete we're doing. And we're going to have time later to talk about this as well if you want, okay?

Bias and Discrimination

So first one, bias and discrimination. So we all know the potentials of AI, but when you start using it in real, and especially if we're looking at public sector, this is very complicated.

We have cases in Belgium where they use a tool to matchmake unemployment. They took it out because people were scared of how it was happening. So sometimes it's a problem.

We have the recent Dutch scandal that got the prime minister fired because of an algorithm they were using for giving out grants, which turned out the algorithm was working badly and giving grants to the people that should not get them. We have bias with COVID in America where black populations were discriminated because the data sets were biased towards the white community.

So we have all these situations, we all know this, and this is very serious as a city. Not that, by the way, Porto is not doing any of this at the moment, but we don't want to do it like this, okay?

Privacy and Digital Rights

So this is one of the things we are looking at. How can we make this more protected in the city?

So yeah, that was it. So as we talked before, the EOI, Artificial Intelligence Act, it's a really good step in that direction. It's very strict.

So it's going to stop a lot of bicycle innovations. A lot of tech companies are going to complain, but it does protect us citizens on this scenario.

Digital Transparency

And we're looking at things I'm going to show later on algorithm registers. So where you can sit and see what algorithms are working and how do they work in sites.

And we have very strict policies on not accepting black box unknown algorithms. So I know Neural networks are amazing, but if we cannot understand how a decision was made, we are not going to put it in a situation for a citizen.

So that's not possible. We need to figure out what happened and why it happened.

So we are waiting a bit more for the explainable AI functionings to come into play so we can use these algorithms in real life.

For privacy and digital rights, I'm not sure if you are very aware, but there was a case in Porto where a robot operated by a company that does remote control, so you can control a robot from your computer, was walking in the city. No one knew about this. No one told this. It was stopped immediately as they found out.

Of course, we employ GDPR every hour we can in our services.

And also, cities in Valongo and Porto are looking at regulation for drones, which is also another issue.

just a heads up i'll do this again later we are going to do a hacker city event on the 11th of december on this topic we're going to share with you our wi-fi data which you can then play a bit around and see how you could do analytics with privacy preserving so how do you do analytics without invading people's privacy the data set is safe completely anonymized but you can play around to see how would an official do that so if you want to play your team you can get money out of this so this is some free publicity for the event i'll have a qr code at the end

1For digital transparency, which is a focus we have this year, because we're celebrating digital rights this year, we're looking at what Amsterdam, Eindhoven, Helsinki, and other cities have done, which is this AI algorithm register, which we try to implement next year, which is a page where you can see all the AI algorithms work in the city of Porto, and how is the algorithm working. Currently, we don't have many. The one that Jose is talking is a new acquisition that would eventually be here, and now we are working with these.

And what we launched this year was a city digital systems information platform where you can check which sensors are working in the city of Porto and how they are taking away your data, let's say, and the data chain, so is it used or not. And I think this video would play while I'm talking. So I don't know if you notice around the city, you're gonna see it.

If you walk around, you're gonna see these little posters with QR codes where you can check with your phone and they're gonna take you to a page where it shows you all the data, all the sensors, and what data are they taking and for what purpose. And this is being tested now as to see how the citizens feel about it, if they like it, if they don't like it. Our first discovery is that most citizens do not know that there are sensors. They don't know what they do, and they are happy to know more about it.

But yeah, feel free to see if there are other cities to implement and play with it. It's going to, in principle, be launched everywhere in the city where there are sensors from the city. For also the purpose, which is sometimes the sensors are not from the city and they should not be there. So this also helps us to understand who is using the city for what purpose. So this is for cameras, air quality, and I don't want to pull much, we can talk about it later as well if you want to talk with us.

Community Involvement in Tech Solutions

For digital inclusion, which is also another topic we have, I'm going to invite Adip to talk about community city, to focus on how can we get these technologies to reach the, let's say, less digital communities in our city, because the potential of AI is big, but it's also bigger to widen the gap between the people that already don't know about technology to even be more, let's say, sidelined from society. So thank you and I will carry on where we left off.

So this is a CommuniCity project. I'm gonna run this video a couple of times so you could look at it and look at me as well. I will try not to walk. I will try to stay here for the camera as well.

I forgot. So CommuniCity is a European funded project. It's a three-year project. We have around 12 partners with us.

But the core thing of this project is including citizens in any tech solutions to solve their challenges and their needs. So we had three rounds of open calls and with each open call we call tech companies in order to present their solutions, whether they are AI or other digital solutions. and they should answer the needs of the citizens and for that we do some citizen-centric or human-centric workshops and innovation in order to identify their challenges their needs and how we can solve them with these digital solutions and then afterwards we launch these calls we have 12 500 euros for each pilot and each tech company have around five to six months in order to implement their solution by co-creating with these citizens.

In Porto, we are covering one of the social neighborhoods, which is Campanão. We have several social neighborhoods over there. And for that, we have elderly people, we have also healthcare institutions, we have schools, some gardens. So we have several things over there and we had several needs identified.

So in the first and the second round, I could give you some examples that happened here in Porto, but also we had three cities with us. It was Amsterdam and Helsinki. In the second round, we had seven cities. And now we are running the third round, which we have 17 cities.

CommuniCity Project

But in general, our solutions were we used AI here in Porto. first of all we had in this association of policy which is a bit visit for Portuguese people so we put over there some sensors to track people with cerebral palsy and what we did is train these sensors or the algorithm of the sensors on to track these people and see their movement because it's different than a normal people's movement and since we don't have healthcare providers with them 24 7 so these sensors could also give some notifications for the healthcare providers on how if they should go there to the rooms to take care of them if we have dangerous situations and these sensors they don't have cameras they don't have any invasive kind of solution so it was just tracking the movement of the people if they fell down they give notifications and that's it but it was one of the successful pilots so this is

actually one of the sensors. So it exceeded its duration and actually the association stayed with the solution up until now. So it's running for like one year up until now.

For example, one thing that happened in Amsterdam It was, we created some chatbots, some avatars, and co-creating them with the people. So they made their looks. So for example, if they are Arabs, they had like veils. If they were a woman, they had darker skin, they had whiter skin, depending on the language they are speaking.

And they were there just because their marginalized communities where the challenge were implemented, they were immigrants and they wanted to speak to them in order to facilitate some bureaucratic stuff or when they want to apply for some documentation. So with that, they had these chatbots where they had live translation using AI and it helped people to feel safe to communicate with this chatbot because it looks or feels like they are talking to a person as well so with all of that we are trying to bring these technologies and their solutions closer to the people by doing the solution with them and not just for them it's not a top-down model but it's a down-up model because they are identifying their challenges they are identifying the solutions they want to work with and they are working directly with the tech companies

So we are getting them closer. But as well, for example, now in the third open call in Porto, what we would like to do is, I will just play the video again.

just not to have a black screen behind me, is trying to get AI closer to the people. So we are searching for solutions. And now I think we have around five solutions.

Bringing AI Closer to Marginalized Groups

We are trying to see which is the best one in order to teach these marginalized groups, whether they are elderly people or just people living in social housing, how they could benefit from AI because we as students or tech providers or people working in the digital world we have direct access to AI but these people are even they don't know how to use their cell phones or They don't even have smartphones or tablets or computers.

So we want to get this technology even closer to them. We want to teach them how to use these devices by leveraging AI as well on their side.

So in general, this is it.

Conclusion

And based on your question, we are also considering the AI Act in our work. So we are also giving some knowledge about the AI Act in our work.

Inside these pilots from five to six months, we are trying to convince everybody to test and do whatever they want because it's innovation and we have a controlled environment as well and we control the people that we are working with. But at the same time, we are already teaching them that, OK, if you want to carry on your solution to the market, you need to be careful of your solution, if it's compliant, if it's not compliant with the AI Act.

So this is it. And I think I'm going to end.