The following are the outputs of the captioning taken during an IGF intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.
***
>> PRATEEK SIBAL: Okay. You are in the same region.
>> MARCO: Good morning, ladies and gentlemen, our session will start very soon. Let's give it another minute.
Yeah. That is the thing. The sound is very computerized, Marco.
>> MARCO: Sorry?
>> PRATEEK SIBAL: You know in the 90s we spoke into a computer and it generated a voice. It sounds like that. Can you try speaking again?
>> MARCO: Hello. Could you hear me well? Maybe the technician in the backstage can help.
>> PRATEEK SIBAL: We can hear you. We will manage unless the technicians can fix it otherwise, we can hear you.
>> MARCO: Okay. Good to see you. Ladies and gentlemen welcome to the session of unlocking the power of AI across the public sector, media, and judiciary: A UNESCO‑IRCAI online repository of AI tools.
I will be the onsite moderator for this session. My colleague Prateek Sibal who is a Programme Specialist in the same sector will be the online moderator. This session will consist of two parts. The first part will discuss AI's impact and potential in these three sectors.
In the second part we are delighted to give a world view of repository online with stakeholders in the three sectors and you will have an opportunity to ask questions to panelists at the end of the session. We have (?) And UNESCO joining us onsite and online and first and foremost we honor to have Mr. Tawfik Jelassi here with us and would like to invite him to the stage to deliver opening remarks. Mr. Jelassi, the floor is yours.
>> TAWFIK JELASSI: Thank you very much, Marco. Distinguished participants, colleagues and friends, thank you for coming to this session and for UNESCO organizing and unlocking the power of AI across the public sector in particular. The media and the judiciary and the session as it was mentioned is jointly organized by UNESCO and its Category 2 center for the International Resource Center on AI.
We will be delighted to present to you the new initiative, which was jointly developed by us, UNESCO and IRCAI and looking at impact of AI on them and the public sector and the media and the judiciary and we all know the impact of potential of AI not only in terms of boosting efficiency, but also opening new opportunities for value creation and the impact of it on education. You may know that a couple of years ago, UNESCO was the lead partner of the United Nations in transforming the education summit that looked at the impact of frontier technologies including artificial intelligence and generative AI.
I will talk to you about health and planning and other sectors, which, of course, are relevant to any society, to any country.
To support this work, UNESCO has developed in the context of the UN broadband commission for sustainable development CAV4 it develops a competency framework on AI and digital transformation, a framework that is aimed at improving the capacity, the competencies of fewer servants in the context of the public sector.
In June of next year, UNESCO will be hosting in Paris a global conference on AI and digital transformation in the public sector, again, putting the folks very much on capacity‑building and capacity‑development. Why are we doing this?
Based on our needs assessment survey done in Africa but also done in small states and small islands and developing states in the countries, the No. 1 big question we will see is capacity building and capacity development. We see no country can successfully implement digital transformation offering better digital services to its citizens without civil servants having necessary skillsets and competencies to do that.
Let me say also that UNESCO has been very much involved in combating this information and misinformation, hate speech and other online content. This is an initiative that would have started three years ago.
We are continuing with that, and we believe that AI not only is producing this information but in amplifying it and also represents a new challenge that we all need to tackle.
In 2023 last year, UNESCO posted 30 media organizations in central West Africa and also in southeast Asia in adopting AI tools for reporting on climate change.
So, the issue of climate change, the environmental crisis, which is one of the biggest challenges, is also at the heart of the UNESCO programme regulation, in particular to media and the role of the media in tackling the environmental issues.
Next year, we will be focusing on our commitment to this subject. UNESCO was last month invited by the G‑20 to be the secretariat for G‑20 global initial on information integrity for climate change and UNESCO will have the global fund for integrity on climate change.
This is a new global initiative that we are all being asked to lead it and to be in charge of it on behalf of the G‑20 countries and this re‑enforces our work on this specific topic.
Let me say that since we talked today about the judiciary, UNESCO now for 12 years running has trained more than 36,000 judges, prosecutors and judiciary operators in 160 countries worldwide.
First, the training focused on international standards on freedom of expression and the safety and the protection of generalists. But for 2 years, starting 2 years ago, we added another programme, which is data of AI on the floor and so far, we have trained more than 8,000 judges and prosecutors on the impact of artificial intelligence of the judiciary.
This is something that is very important for us because we believe that judges, judiciary operators hope to be aware of the potential but also the hazards and the risks that new and emerging technology can have on their work.
And in this context, I'm pleased to say that the constitution of Colombia recently highlighted this initiative of UNESCO and related to AI as a model for integrating AI into the judicial system. We are delighted by that, and we hope that you can take this pioneering experience in Colombia and hope to take it to other countries around the globe.
And for that, we, of course, need international cooperation. You have here honorable judge from Tanzania.
We have other members of judiciary in the audience, online and in the room and talking about collective efforts and partnership, it is not just UNESCO alone but partners especially from the judiciary so we can be relevant and hopefully impactful.
At the end of the day, that is what really matters.
Let me also say IRCAI and UNESCO are developing a global repository of ethical AI tools for the public sector and the judiciary and the media.
We hear more about this initiative today by the ‑‑ some of the speakers in this panel.
Let me maybe stop here because this ‑‑ these are meant to be just opening remarks.
Again, to thank you for being here and to say that the work of UNESCO on this technology has started, I would say, some 6 years ago, 2018, when we started working towards what has become in 2021 the first Global Normative Standard on the ethics of artificial intelligence.
This is a landmark achievement. The UNESCO recommendation on the ethics of AI voted by 193 countries back in 2021 before the emergence of ChatGPT and generative AI.
This is what UNESCO; I say it in a humble way. UNESCO is known to be a laboratory of ideas trying to have some foresight and trying to be a bit proactive and thank you so much. And I think this ‑‑ this example shows, again, that without Member States, we have stakeholders from Civil Society, from academia, from research that work with technical community and we are able to put together and we worked on this from 2018 to 2020 and then 2021 it was called for a vote and all Member States of UNESCO, 193 countries approved the UNESCO recommendation on the ethics of AI. We know today this is a big issue all over the world.
How can we make AI use, be responsible, be ethical so we can benefit from it and yesterday the UN Secretary General in opening remarks here said that technology is here to serve humanity not the other way around and technology to serve humanity technology has to be used in responsible and ethical and human‑centered way. This for us is the compass going forward.
Let me conclude by thank you and also the government and ethics professionals and developers for collaboration with us to harness AI for public good enjoy the session. Have a good day.
>> MARCO: Thank you have he much for your inspiring future words and AI can harness for the public good and resonates for the theme with this year's IGF. Lots of themes for good thoughts. I would like to invite the distinguished panelists to the stage.
We have with us in person Ms. Amrita Choudhury and Honorable judge Eliamani Laltaika from Tanzania.
Please join me on stage.
The two panelists are unfortunately missing out on this grand venue are Mr. Henrik Ahrens, the French media development agency and Mr. Davor Orlic, Chief Operating Officer at the International Research Centre on Artificial Intelligence, IRCAI.
Welcome to our four panelists.
So, let's begin with the public sector. Amrita, you have extensive knowledge of Internet governance and digital innovation in governance in Civil Society in India and also in South Asia more broadly how do you view AI's existing impact and potential in the public sector.
>> AMRITA CHOUDHURY: So, thank you for having me here. For a developing country like India or any of the South Asian countries wanting to leapfrog using digital technology, I think AI is a lessee and that is how the governments also consider it.
For example, the previous speaker was talking about climate change, agriculture, et cetera. These technologies can help increasing output and taking care of climate change, et cetera.
That is how the governments have been using it.
Since this ‑‑ you know, the discussion is more on public sector judiciary, et cetera, India has been using AI in the judiciary also.
Certain facts, which I would like to share is like the Supreme Court of India actually translated 36,271 judgments into Hindi. India has official 22 languages. Normally it were ‑‑ it used to be in English and the local courts would use their own languages and translated into Hindi. 17,142 translates in 16 regional languages using AI. Primarily it is used not only for translation but also for document research, process automation, prioritization or categorization.
There is a lot of cases looking at case histories to kind of find references as something where it is being used.
And even also for transcribing the oral arguments in the constitution bench.
It started being used from February 2023.
In terms of other public sectors, it is ‑‑ you know, the digital public infrastructure is something that India is promoting especially the other that is the unified identity number and the unified payment interface, which we normally use mobile because 90% of India uses mobile. Any kind of payment even through a seller can be done through the mobile platforms.
Those are certain places where AI is being used.
One of the things ‑‑ misinformation and fake news is something that is a challenge. AI is being used. It has to still go a long way, primarily because local languages, context, it is something that machines still have to take up.
I think the frameworks, et cetera, which UNESCO has been working on and hopefully will be releasing will help. Another thing where AI has been started to be used recently is for SPAM and fake messages.
There is a lot of financial frauds that are happening at India at this point in time. It is huge and even deep fakes and everything. Telecom companies start today use AI to label and tell you this is maybe potential SPAM. Lots of messages are coming, even platforms like WhatsApp and Nectar is used a lot and using AI to alert users.
Everyone has been pushed into technology, but they don't know how to use it responsibly. I will stop at this in case you have anything else. I will be happy to add.
>> MARCO: Thank you, Amrita. It is fascinating to see how AI is offering governance and Civic engagement and impact on it in the media sector. Henrik you have decades‑long experience in media development and in the Middle East region. Could you tell us how journalism and media operate in general are still the same as AI is increasingly used in newsrooms?
>> HENRIK AHRENS: Thanks, Marco. First of all, decades is overexaggerating my age. A decade and a half at best. So please, I ‑‑ I will not be able to give you like a general perspective on media given my role here as Director of the programme that is covering four countries in the Middle East notably Jordan, Iraq, and Palestine, and I would like to share. I'm not a techy guy and working with media and journalistic background and would like to share a couple observations that might be very interesting for you.
Calip, the term means proximity and meaning the objective of our programme is to bring the media closer to the audience to remain relevant.
So, one of the main questions we continue there is how to you ensure relevance especially in light of increasing influx of AI‑generated content.
How do you develop formulas that are up to date and normative and how do you compete and how do you ‑‑ how are you able to increase the pace of your production?
So we carry out regular small surveys with our partners in the four countries and as a matter of fact, as a result, we find out that the vast majority, like around 80 to 90% are like in an experimenting phase, meaning there is no internal policies and no approaches. However, everybody seems to be aware there are tools and challenges available. Talking about the challenges I will talk about a couple positive examples.
For instance, an Amman‑based and Jordan‑based podcast organization producer created a text to speech technology called (?) That is one of the best working in Arabic language and facilitate production of podcast and it is really something outstanding and especially since digital transformation in Arabic includes the ‑‑ another layout ‑‑ added layer to the challenges and the language and they managed to come up with a state of the art impressive tool and also media Iraq‑based radio station that operates a model in Neva a quite challenging context and they have a techy background and managed to come up with a self‑made technology using AI in order to analyze content, social media content that is sort of AI‑related and AI social media listening tool that is prompting editorial decision‑making process and able to capture on‑time and real‑time behavior and audience engagement relating to specific topics that drives decision making and makes at the end of the day better coverage and more coverage and another tool that is very well‑working and has led to an increased audience and reached the radio station.
The other media outlets I want to name another one (?) That is an online platform that is basically embracing AI‑generated content and are inviting content creators and journalists and producers to come up with creative comments to AI for innovation and policy of that is clear to operate to journalistic coverage. Of course, in that case, it is an invitation for an experiment to come up with something like that.
At the end of the day, I already mentioned the statistics and the majority of the media we are working with which are independent and small to and best sized media operations and are rather seeing a variety of challenges.
There is at this moment in all of the ethical journalism and ethics challenges and technical challenges and human financial resources and, of course, internal and external policies and internal policies ethics related vast majority of the media realizes they receive from especially freelance content and journalists sometimes AI‑generated content. It is not making transparent and lots of media have increased awareness for the risk, we will say.
There is not a real policy and not a clear approach about what is acceptable and what is not and there is a common denominator that you can't just deny it and how do you basically enable media to have a journalist and use AI without being and ensure for now that is content and unique and timely and relevant and research that is a very big challenge.
Whereas there is a debate, there is not a clear point of reference at this state.
I mentioned already technically the Arabic language.
You are all aware about the debate of digital divide and increased need for digital transformation and especially in the context. I'm operating in the Middle East and Arabic language makes it even more difficult.
That means also let's say we have challenges and information and AI information and huge challenge for media landscape and discourse in the region and same time Arabic language opens up a space, an opportunity, to close it in a progressive and constructive way, meaning the impact of let's say deep fakes we have in English or myself I'm German and German language content is not to the same extent.
Of course, there is a huge extent of disinformation and constructive narratives to undermine open and diverse pluralistic discourses in all of the countries I work in. There is still way to develop AI tools to come up with a more impactful, tech‑based culturing alternative discourse.
I mean, at the end of the day, for all of the media we are working with, they are independent. The ‑‑ and the funding. Let ‑‑ the investment content in the regions is very, very difficult, especially for these media outlets and has an impact on resources. You need financial and technical resources to keep abreast of all of the challenges.
A lot of media and those that are wealthier and grown, they are seeking expertise and support from mostly other continent ‑‑ context. It would be very relevant to build up resources here.
I mean, the most resources, especially in Arabic languages and more expensive. It is that creates another burden and challenge for a small media outlet having resources and having resources such as UNESCO tools and repository that is already providing a great context for media. My last point I'm running over the minute. Sorry for this, Marco.
I just want to mention we are of course aware of AI. We are trying to use it to integrate Arabic challenges and opportunities in all layers of our programme and are working on media viability and we are also working on viability and UNESCO integrating here and to inform national policies that can provide guidance and ‑‑ and a tool and a point of reference for reforms that can in the best case improve viability of all media in these countries. Thank you very much.
>> MARCO: Thank you, Henrik, for your points and media and impact is profound and increasing using AI in their operations to help also smaller media outlets capitalize on opportunities while managing risks.
Judge Eliamani could you explain to us how AI is currently used in the judicial system and why stakes are so high.
>> JUSTICE ELIAMANI LALTAIKA: Thank you, Marco. Humbly a little clarification, I'm a judge. I don't represent any judiciary here. I'm not even representing my chief justice. Just in case I say something that does not rain these are my own thoughts and ideas.
Thank you very much. I need to say first that someone was asking me: What does the judiciary media and public sector have in common? Why do you put these three things together?
Huh? Judge, please tell us. I needed to address that before I move on.
That personally, I think it is such a well‑pressed title and initiative. Because the media and judiciary and public sector share at least four things.
The first one and first element that underlines both of them is they are generally not geared towards making profit.
Journalism is not a profit‑making endeavor and neither is the judiciary nor the public sector.
Secondly, there is a very high expectation from the general public on observation of ethics.
If you are a judge or a journalist and you do something wrong, the public (?) Is very high and expect today be above the mark when coming to observing ethical issues.
No. 3 is many of the ‑‑ most of these organizations falling under these three groups are free from the ongoing war between copyright and (?). For example, the judiciary produces a lot of documents, which are not copyrighted.
Everything I write is a judge and it is out there for public consumption and free and available copyright. If someone is training AI tools using judgments, nobody will sue them for violating any copyrights.
I write thousands of pages per year that are all available in Tanzania. We have Tanzania legal repository institute.
Every single judgment I write every day is uploaded on the Internet.
It is free for anyone anywhere in the world to use.
So we don't have way somebody may sue because they have used those judgments to train.
That is why we are seeing so much improvement in terms of ‑‑ of AI or whatever ‑‑ all this generative AI.
Lastly, the three sectors are geared towards empowering the community at large.
Whatever they do, which is wrong and could have catastrophic effects.
For example, the media is regarded as 4th (?). We have three branches of government: Parliamentary, executive, and judiciary.
Media is the oldest and stands on its own and have the power of changing the mentality of people. If disinformation is allowed, you can be sure that so much is at stake.
At the moment in the judiciary, we use AI for the following purposes.
No. 1, we use it for judicial scheduling.
I no longer have to ask someone to write all of the cases that I have to decide.
I have my monthly schedule that is powered by AI.
I can eventually call my phone to know that next week I have this in these cases.
No. 2, this AI is usually in legal research used.
Like I said, we use it wholeheartedly. We are not violating any copyright out there.
If I such ‑‑ a president from the US or UK, I know that of all of the countries including the fellow panelists that explained very well about India, they are copyright‑free. I use them with a clean heart.
AI is usually in judgments. Like I alluded to in another forum, our assistance, every judge is given a legal research assistant.
Our legal research assistants are young men and women who have completed their degrees or master’s and they are imagine straights. They are very good in ICT.
We have seen their output has improved. My legal research assistant brings me an opinion and drafted judgment. I'm like, wow. I am not seeing any gaps.
So, I think that is a very positive way. But also, we are using it as language translation. There is a tool that we go to that actually was adopted from India and very much thank my panelist that that is what India has. They are a trailblazer in most of what happens in the Global South.
We think we are more similar to India than probably the UK and other highly developed countries. Thank you very much. I will stop there for now.
>> MARCO: Thank you for sharing experience and views and impact to justice and how marvelous is that? Such that AI tools must be handled with care to become catalysts for positive changes instead of sources for bias and inequality. But which AI tools are we talking about? Where to find them and how to use them?
So as we come to part 2 of the session, I will pass the mic to Davor Orlic who will give us a demonstration of the better version of online repository that UNESCO is currently developing in partnership with IRCAI and you can have a collaborative initiative and you have the floor now.
>> DAVOR ORLIC: Thank you very much. Good morning. You can hear me, right? Yeah. Let me just try to share my slides.
I should probably also have less slides than more slides.
Just as a ‑‑ as an initial introduction, we would approach by Marco a couple months ago at the IRCAI center and they came up with this problem which was so many things out there. We don't really have them centralized anyway.
The more we started to talk, the more we realized that those things out there would have been AI tools that people actually cannot find. Fortunately, because there is so many of them in so many ways.
On the other hand, the differences are also huge because of different AI techniques for different AI tools and different AI methods being used in different countries and different languages and for different purposes.
This idea was very appealing to us. How do we create some sort of list to make sense for everyone to sort of tap in and use?
The task was quite interesting in the sense that particularly Marco was very much interested in how do we create this list of tools that would have been coming from the global north and Global South and how could we make it work for everyone? So that equity idea was really good.
It was appealing to the team. Maybe if just, I share these. Can you see the slides? Yeah? Is that good?
>> MARCO: Yes.
>> DAVOR ORLIC: On your right‑hand side you have three sort of chunks of content with three different columns that would have been three topics we are looking at that would have been the judiciary public and media and then the second row you would have the radar itself that is a drill‑down tool and guidelines and submission page. We asked not just users but member states to contribute to the repository that would be in English and Spanish is where we would really like to make it even more accessible with even more translations.
Not just the page but the whole content. The third one or layer would have been like a really hardcore drill down into the tools.
If I pop in very simple and multilingual, again, high compliance and very seamless and easy to use.
Low bandwidth was important for us to use and optimize for users with connectivity. I will show you user-experience of the reader needs to be quite quick and understand because this is meant for practitioners across the world and not necessarily from the big outlets, especially the media, but from smaller ones that don't have that much knowledge of AI.
Homepage, again, public sector media and judiciary, as you can see, quite simple and effective and second one again, media and judiciary and the juice, should I say the drill down. Should we have quite an accomplished I think data base where we try to categorize the tools as effectively as we could.
Also making it very, very understandable. Now hopefully you are looking at the public sector media and judiciary and idea is to extend on the topic side.
Finally, again, this would have been a different view and is very modern in the sense of colors. This is state of the art in terms of usability. And that would have been the ‑‑ the visualization on an entry level.
We would have had the intention and the stakeholders and maturity whether it is a company or open source and in which language and whether it is a tool within the company or whether it is a tool within a portfolio of a company.
Then with the hashtags and what actually does it do, and does it analyze or share or interact with you as a tool? We obviously point to the Website and potentially to GitHub, which is where the code is if it is open source.
The idea of the reader is essentially for ‑‑ actually, we don't want people to stay on it.
We actually ‑‑ the success KPI for us is for people to actually leave the repository ‑‑ the reader and repository as soon as possible which means they have found what they need.
So that is successful.
>> MARCO: Davor I was told we have 3 minutes left for the session.
>> DAVOR ORLIC: That is okay. We encourage you once it is out to help us out and populate it more and that is that from my side. I will stop sharing.
>> MARCO: Thank you, Davor, for this. This repository is under development and will soon be open for testing. We invite all of you to provide input.
All of you will be part of what we will actually make it useful.
So, since we are short on time, we don't really have a time for a Q&A session. We might have time for one question if anyone would like to raise a question to a particular panelist. We don't have questions online? No. Okay. So I will ask each of our panelists to give us a 20 to 30 seconds ‑‑ if it is 20 seconds, it will be great. Key messages they want us to take home today. I will start with Amrita.
>> AMRITA CHOUDHURY: Promote as much as possible to all countries, yeah. So people use these great tools which are there rather than try to create it themselves.
>> MARCO: Thank you.
>> JUSTICE ELIAMANI LALTAIKA: Pick up courage and embrace the unknown and create for future generations.
>> MARCO: Henrik?
>> HENRIK AHRENS: I think for all media playing, media to make use of available resources and keep yourself informed and to continue educating and to embrace the change.
>> MARCO: Davor?
>> DAVOR ORLIC: Mine is very simple.
Once the reader is public and available, make sure you forward to your AI experts for them to give us more entries so we can build this database for practitioners across the world to use.
>> MARCO: Thank you. Thank you very much. Now we come to the end of this session. I would like to thank ADG and panelists and those that contributed to the sell and repository and the process. Thank you to all of the audience for joining us today and IGF secretariat and the Host Country for hosting this session. Thank you very much. Let's give this round of speakers a round of applause. Thank you.
>> PRATEEK SIBAL: Thank you colleagues, Davor and Henrik. Happy holidays. See you.
>> DAVOR ORLIC: Bye.
>> HENRIK AHRENS: Bye.