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.
***
>> ANNOUNCER: Please welcome to the stage the moderators, Allison Gilwald, ICT Research Africa, civil society Africa and Pablo Hinojosa, independent GRULAC.
>> PABLO HINOJOSA: Hello. Welcome. It's not a big crowd, but it's a good crowd. So that's important, and we also have not the ones present here, but online, and hello to all of them.
This is also for preservation because we will be recorded and I think it will be an important dialogue. I want to welcome our fellow panelists, and start a very good main session on emerging technologies and human rights.
>> ALLISON GILWALD: Thank you. Let's call them in.
Well, this is it. Show time. I think time mostly intimidated by the Resolution of that camera. It's really Hi-fi. My name is Pablo Hinojosa. I'm here for the Marconi Society waving the flag of Internet resilience. I consider myself a product of generation WSIS. And for the purpose of this session, I will be wearing the badge, now it's a digital badge -- hi, Julian, it's good to see you in line -- of teen technology. And I have the absolute pleasure of co-moderating with Allison Gilwald from ICT Research Africa who brings the wisdom and depth of teen rights, teen technology, teen rights, but for the record, we are not in opposite sides at all.
I think the very point of this session is that technology and rights must be of the same group working together to shape a future that is inclusive and people oriented.
I think we all in the preparatory process agreed with that at least, but hopefully we won't agree on everything, just on some things and make this a good session.
So we will have a difficult task of moderating this group, an incredible panel of speakers. For the session structure, we have 90 minutes, and we are going to split it in three parts. I'm not sure how many of you have read the brief of the session. It consists of three questions about emerging technologies and human rights. And we will rotate speakers across those blocks Allison Gilwald will moderate the first and second, and I will moderate the third.
We encourage audience input in each of the blocks. My understanding from the producers is that there are some crosses in the hallways where you can pick up a microphone and you will be on, but please do not derail the discussion, bring and be part of the dialogue and be very fast. We appreciate that very much.
We will be closing with reflections and forward looking messages, and that's on the logistical side. On the panelist side, I am really honored to be working with all of you.
To my left we have Pierre Bonis, CEO of AFNIC and he is also team technology, I think. I don't want to categorize everyone. I think we all should merge and converge.
So for those that don't know, AFNIC is the CCT of the registry, and he brings a wealth of really good points in the session.
Next, we have Rodrigo Goni. He is an American friend. He is the head of the future of the commission. Rodrigo is President of the Commission of the Future of Latin America. He participates with us as a parliamentarian. This is a really important contribution. Together with Rodrigo, we have the Executive Director of ICT for Change, a well-known entity in the IGF ecosphere, and I think she is on the team's rights working together. It's an honour.
Peggy is super important your presence here. Thank you for this. She is the Director of the UN High Commission for Human Rights. And your expertise will be welcome to help us frame and give adequate language to these discussions.
Alexandra Walden from Google, she will be bringing perspectives from the global policy side, and we are also very thankful. You are probably in a difficult spot, and that's part of the point of this, but we are really welcoming sort of the conversation.
Julian, I met him online in the preparatory process. It's really an honour to have you here. You helped a lot with the framing. You were the one that says these are not opposing teams, so it's a credit to Julian for that.
He is in Prague. Are you in Prague?
>> JULIAN THESEIRA: Yes, I am in Prague for my Ph.D.
>> PABLO HINOJOSA: Julian is at the Center for AI and Digital Policy, and I can only speak for the quality of your research. It's impressive. Please look for him online. So that's the panel.
So we don't need to introduce the panel each time that we go ahead, so I think I got rid of all of the logistics, and let's get onto the substance.
What is this about, Allison? Let's frame the conversation?
>> ALLISON GILWALD: Thank you so much, and thanks for this great panel joining us today and what's been so far. So perhaps just to start by saying I'm from team people, so I'm not from team technology, and I think a number of people, people on the panel, but also in the room have been working for some time on trying to extend more sort of techno legal understandings of rights and digital rights to more socioeconomic technical legal but also socioeconomic and talking about rights in a very people-centred way. And that's largely because conservatively sort of two billion, 2.6 billion, but more accurately about 4 billion people remain offline. And often these discussions around digital rights exclude significant parts of the world's population.
And even in those jurisdictions where there is some resorting to some kind of rights framework, people very often don't have the capabilities to exercise those rights. So even though they exist on paper, how do we realize it? How do we make them more practical?
In terms of a number of global forums and a number of alliances, global justice forums, we work with Anita on and a number of other international forums, Global Partnership on Artificial Intelligence has been dealing with issues of extending notions of data governance, data justice. And just answering or asking the kind of question, can you have rights-preserving, ethical governance frameworks that still produce unjust outcomes? I think we only have to look across the globe to know that's the case. We have Universal Declaration of Rights. We have these frameworks, the Human Rights Council that we normally appeal to, but in practice these rights are not exercised by a large majority of people so how can we do that?
I think part of the critique has been, of course, emphasizing the importance of first generation of rights, of fundamental rights, of rights of privacy, freedom of expression, access to information, absolutely critical, but that those aren't sufficient for the kinds of redress that we need to, around digital inequality and rights inequality more fundamentally, because I think for many of us digital access is an enabler of the, of exercising one's rights in a content-free world.
So I think many of us have been working practically, of course, feeding into the Global Digital Compact and the WSIS+20 review considering, you know, after 20 years, after the COVID crisis, et cetera, how far have we come? And how big this gap is?
So I think some of the things is just moving beyond those first generation rights which have been practiced in a very kind of individualized way so that individual rights have taken preference over the collective rights or public interest rights. I think we saw this particularly around COVID.
And then also kind of the more systematic inequalities and injustices and rights abuses that one sees because of structural inequalities, so that sort of thing. So I think moving, extending because it's not that those individualized rights aren't important, but extending those rights to look at the collective implications of that.
And then as I said, extending that if one is really concerned about redressing the current inequalities we see from a rights perspective is shifting that to second and third generation rights, so to economic and environmental rights that would allow people not only, you know, it's not only a compliance framework for rights, but it's actually the kind of enabling framework that you would need for governance of technologies to ensure that people can exercise those rights, enjoy their economic opportunities that are offered by the digital economy and data society.
And I think those are where the challenges are coming now in regulation. Really shifting, as I say, not doing it at all, but extending the personal protection kind of rights frameworks, digital rights frameworks to non-personal powers, power asymmetries we see in the data economy and data society.
That's the framework. And with that I think this is a very good scene to start a conversation, and Allison, we have the first block, the first policy question that is about how can we use existing international Human Rights standards and instruments by different stakeholders to be improved, to prevent violations and abuses, and how to make a digital environment that is well designed and people oriented.
So that's the first block and I will leave you to moderate it.
>> ALLISON GILWALD: Thank you so much. So in the first section, we have Peggy, Pierre and Julian online to discuss this question that Pablo has articulated for us focusing on these mechanisms of preventing and addressing Human Rights violations and abuses in the digital environment, but also how we can, I guess, maybe proactively or precautionarily anticipate some of these violations and how we can mitigate those effects. So Peggy, if you would start us off.
You've obvious are obviously in a strong position to speak about Human Rights impacts and precautionary tools that exist. Please do share those with us.
>> PEGGY HICKS: Great. Really happy to be with you today and with those gathered in the room.
I think it's fair to say in a big room like this it's, we talk about sometimes elephants being in the room, but there is an elephant in the room as well, which is why are you all here? Why are we still talking about human rights. And I think Allison has put it on the table quite well that there is a real tension around what human rights mean in the digital space, and I think those of you that have been following it have recognized that it's actually gotten the conversation has changed quite a bit in the last year, year and a half. And we really need to take that on this panel, I think. We have to talk about the fact, like, why is Human Rights relevant? What does it mean in the technology space? And most importantly, how does it relate to the need for innovation and succeeding in a competitive world around development of digital technology.
And I think the most important thing I want to bring to that conversation is the idea that really getting human rights right is not a hurdle to innovation and development of digital technology, but it's actually an asset. It's ultimately the companies and businesses and Governments that figure out how to do digital in a rights respecting way that will see the greatest advances that benefit people, and this goes to Allison's point as well about how we ensure that the real benefits of Artificial Intelligence and digital technology are there for everyone across the full set of rights.
It's interesting you talked about generations when, of course, economic and social rights were there at the creation. We have a whole covenant on them, which is just as old as the covenant on civil and political rights, but I think you are right in terms of thinking have we really fulfilled it in the way that we need to. And why I think the Human Rights framework is so relevant is that we also, as I said, we live in a very fragmented and polarized world, and the human rights framework is still there as a resource that is universal.
It has been agreed by all of the Member States of the UN, and they have all committed to basic principles that guide how we develop and deploy and use technology in a broad sense. And relying on it as a such stone that we can all agree on and bring it into the conversation, I think, is very important.
So human rights principles can both help us define what's at stake, what are the things we need to preserve when we are doing this, why are we doing it? What's AI for? What do we really need it in? Do we need it to better achieve health goals? Do we need it to better ensure education is available? Do we need it so we can have better translation within our panel and put some better spin on how we are all talking to each other?
But also it can help us navigate the solutions, because we know that things do come into conflict in the space, and we see it all of the time in terms of what it means to develop technology that can serve people, but can also ultimately undermine some critical rights.
And the Human Rights framework has been worked on and developed in a course of different issues in a way that allows us to transverse that. It's not that one right trumps another, but that we do have to figure out how we can achieve the full set of rights in an effective way. And we are working on a variety of things I can come back to later that will help us to traverse that space including how we do human rights due diligence around digital technology.
But the three key ingredients I'll leave you with is transparency, so this means we need to know what's happening, we need to understand and have good reporting on how technology is being rolled out and use and how human rights is being evaluated in that context. We need participation, that means that all of the stakeholder voices need to be at the table. Really glad to be at IGF where that's the case, but we need to make sure it's happening in policy conversations around digital tech and AI everywhere, not just because it's the right thing to do, but because if we don't bring in civil society, if we don't have that expertise in the room, we are going to miss a lot of the things we need to know to deliver human rights and technology in the way that we want.
And finally, we need oversight and accountability. It's not enough to roll it out, we need to know how it works for people and there need to be mechanisms to ensure that we are following and learning from what doesn't work and try to improve and ensure we are delivering better results as we go forward.
>> ALLISON GILWALD: Thank you so much Peggy. Just as a quick follow-up, could you speak about, you have spoken about accountability and transparency and participation critically in this forum, but could you speak a little bit about some of the kind of institutional challenges that are there that siloed attention to these rights, frameworks, within a legal framework needing to be extended to these much more kind of dynamic agile institutions we need for the environment we are in.
>> PEGGY HICKS: The greatest challenge is inequality and discrimination within the space, and I was talking to someone earlier of how it occurs at so many different levels. We know that within the tools we are building, we are often building in bias because we are using data sets that have bias within them. And that's sort of the starting point for the conversation. We are also super concerned about whether or not investment in AI for good is what it needs to be and whether or not we are too driven by what can sell in a marketplace as opposed to what will deliver the most value to the people that need it most.
A third level of inequality is around how do we make sure that the benefits of digital technology reach the people who need them most? So the vulnerable people, the marginalized people who really are probably going to be the last to get on board the digital tech and AI band wagon but may actually be the people who could benefit the most and probably the most cheaply at the very outset.
And then finally, of course, we are seeing a huge growth in inequalities among states. You have to look at where are the data centers? Where is the compute? What language is this being done in? To realize if it's not managed in a better way, we will see an increase in global inequality in a variety of ways. So that's one of the big challenges we need to tackle.
>> ALLISON GILWALD: Pierre, I think that provides you with a useful rights framework to go and talk to us a little bit about on the more technical side about some of the infrastructure and access aspects of the technology as an enabler of human rights?
>> PIERRE BONIS: Thank you very much.
I am going to speak in French. So my warm salutations to the interpreters. I hope they will not be replaced by AI, at least not for now.
And in human rights we also have the possibility to speak out and to listen to people in their own language. I would like to speak here about not something that is an emerging technology, but about underlying infrastructure of the Internet that help these emerging technologies to emerge namely.
And that's exactly what's happening. We have this dialogue between human rights, technology on the other hand, and the new technologies on top of that, and it's always good, I believe, to remind everyone that from the very first rollout of the Internet with its own internal character that Peggy mentioned, transparency of the protocols, multistakeholders governance, these qualities through access to population will help these people to exercise their rights.
So it's not a solution for everything, that is true, but access to knowledge, to education, to law, to what is written in the law, the word of the legislation, the ability to upskill and gain more skills. Everything is potentially reinforced by the Internet. I am not saying that Internet is going to solve all of our problems.
It's not just a question of being connected and all of your problems disappear, but it is quite clear that for the last 30 years, the strengthening of these rights and the efficiency have been supported by the Internet.
A point of clarification, one of the rights which is probably most violated today is the right of access to the Internet. Several billions of people do not have access to the Internet, and Internet has a possibility to make the rights efficient and it means that some of the world population is in a very complicated situation. We have spoken about these people for many years, and the longer we speak of it and the most violent it is for these people who are not connected.
They are becoming more and more excluded as time passes. That's just one point I wanted to make. Another point very briefly, the underlying technology of the Internet also has equality. Is it its neutrality. And in our debate between technology and rights, I think we need to remember that the actors of these technologies, that is the one operating them, the critical infrastructure, operators, the ones who are managing these technologies, they need to stay neutral. And there is attention between this requirement of neutrality, and respecting human rights. Well, between these two, we need to find a path and navigate it, and if I am completely neutral, well, Human Rights are not neutral. It's very important, it's fundamental, but it's not neutral.
So we need to find this balance between protocol and pipelines neutrality, and how to take into account these rights and their materiality. Thank you.
>> ALLISON GILWALD: Thank you so much, Pierre. You have spoken about, you speak about the underlying infrastructure because that's so important, the Internet infrastructure that enables the additional layers of all of these technologies. And I think you have made the point very strongly how lack of access to this really prevents significant numbers of people, we can debate the numbers, from exercising these Human Rights, these normative frameworks that have informed the Internet's rollout in this neutral way.
But I think just speaking about how this does affect these more advanced technologies. The lack of representation in these big social networking data sets by people who never come online, inability to financially transact, et cetera, are all affecting the data sets used for these advanced technologies, the big data sets being used for these advanced technologies. So you have an amplification of these exclusions and rights, and I know you have been thinking about this even beyond AI and in terms of quantum and various other things, but perhaps to make the linkages because I think people often think about Internet connectivity as a kind of black and white thing, but it's become the digital inequality is no longer just a divide around connectivity. It's about what you can do, what you have the resources to do once you have that connectivity, how you can exercise your rights and use ChatGPT or produce an alternative technology. Can you just talk to us a little bit about that?
>> PIERRE BONIS: Yes, indeed. It's very important, and this was said earlier by Peggy when she was talking about data localization.
The digital divide and the access divide is also seen on the ground in different regions and countries. The produced data that will be used to access these rights to inform people are not maintained locally. So they go somewhere else, and there you can only hope that when these data are given back to you, it will be done in a neutral way, in an efficient way, and indeed potentially you can be dispossessed of your cultural heritage, of your economy because this infrastructure is not just pipelines. It's also data centers. As we said, this infrastructure needs to be localized everywhere in the world, and not concentrated in specific countries.
So the topic of the AI is not such recent topic. We talked a few years ago about big data at the IGF. We were talking not about AI, but it was the same thing basically.
For the last ten years, we are trying to hire data analysts and if you are in a country where you have no data, there is no point hiring data analysts so it is a vicious circle. It is still a problem that has been rampant for the last 25 years, and things are progressing, that is true, but it's not something that is behind us.
And the more Internet and the new technologies such as AI, et cetera, are bringing good things, the more this divide is absolutely scandalous.
>> ALLISON GILWALD: Thank you so much for this, and I think that takes us to Julian, and you are going to talk to us about human rights impacts, particularly on AI systems. So I think a good segue from our input from Pierre. But also bridging the gaps that we've got on some of the standards and implementation that's affecting the ability to safeguard rights.
So if you could just come in on that, please.
>> JULIAN THESEIRA: Yes, thank you Allison, and thank you very much to the organizers for having me on this panel. So at the Center for AI and digital policy we are a big supporter of international frameworks on AI and AI governance and AI policy that also support fundamental rights, and so one thing we often highlight is there already are such frameworks out there such as notably, for example, the recommendation on the ethics of AI that has been adopted by all 194 UNESCO Member States, so not all countries in the world have agreed to it.
And UNESCO has developed certain tools to help countries to implement these recommendations, so notably the assessment methodology, to assess their national frameworks and ecosystems about countries, about whether or not they are ready to apply the necessary recommendations and then there is also the ethical Impact Assessment tool that can be used to assess the impact on rights across the AI life cycle.
And at the centre of AI digital policy, we are a proponent as well for rights AI impacts assessment and as we see in certain jurisdictions like the EU, like the EU Act mandates assessments for AI systems and there are countries like the Netherlands where there is are assessments mandated by law for Algeria systems. And then in terms of how things can be improved and building capacity, so, again, there are also existing efforts that are ongoing at the same time as this, for example, as the same time as this Internet Governance Forum.
There is also the UNESCO forum on the ethics of AI that's ongoing in Thailand. It's supporting the implementation of the UNESCO recommendations, and our CIDP, Hitchcock, she is in Bangkok training train-the-trainers for civil servants and experts who will go on and train more civil servants in different countries in AI literacy, which will also help to better understand the ethical implications of AI and how they can implement the UNESCO recommendations in their own countries.
>> ALLISON GILWALD: Julian, thanks very much. We are very keen to follow up on the readiness assessments and the impact assessments because a lot of the, for many Developing Countries that's been a significant challenge. We all know it's got to be done, but we often don't have institutional capacity, and I think this is new technology we face some of the same institutional capability challenges so interesting to hear about the examples you have given, but I guess the question is institutionally and on scale, how do we address this?
>> JULIAN THESEIRA: I think UNESCO is making a lot of effort to support various countries and actually the readiness impact assessments have been conducted in countries that would be classified as emerging economies or Developing Countries or there are different terms countries such as Brazil or Kenya, for example, so countries that may not be top of mind when one thinks of like cutting edge AI, per se, but they are interested in understanding what they mean for other countries and societies, and just in working with international stakeholders to better prepare and enter their own domestic frameworks and legal systems can ensure that AI is ethically, and right now the UNESCO forum, for example, as I mentioned there is currently train-the-trainers ongoing, where we are also training some experts and we will then go on to train more civil servants in different countries around the world over the next 12 months, so spreading and sharing their knowledge. So I think this potential could be one way in which capacity building could be spread out and made more sustainable and also equitable.
>> ALLISON GILWALD: Thank you so much. There is so much to discuss. I think we can take one question from the audience before we go to the next panel.
>> PABLO HINOJOSA: I have them because IGF in Benin, hello, thank you for following, is asking what are the biggest challenges in ensuring the security of fundamental rights of technology users in rural areas? I will also mention the question from Timothy Holbern, how can International Human Rights Law be effectively integrated into Internet Governance to ensure humanitarian ICT principles such as protecting essential services from being turned off during conflicts, supporting digital agency for natural persons, addressing challenges, pervasive surveillance, digital consent and access to local remedies via courts while countering adversarial efforts to undermine human centric digital transformation. It's a mouthful of difficult question but those are from the online world. Is it okay if we throw them. Who wants to take them?
>> PEGGY HICKS: Happy to jump in. I think the question about what are the challenges in rural areas has been partially answered in that connectivity is the first one, right, there are so many people who aren't benefiting from the needs that they have and aren't able to access the benefits to digital technology, but I would also throw in there an accompanying problem that doesn't get as much focus is the impact of Internet shutdowns. There are people across the globe who don't have access to the Internet not because it hasn't reached them yet, but because there has been a conscious choice to shut is down. And there are good proposals being put forward around the WSIS+20 process to do better on reporting and monitoring on Internet shutdowns, but I would start with the issue of connectivity as the greatest challenge in rural areas.
And then obviously, it's also adapting to make sure that once you have it, we are using it in the right way to address the needs that people in rural communities have. So I think the challenge of making sure that we are not building systems that are designed with certain audiences in mind is a real problem because the people that are sitting in the rooms and doing the development, it's also a gendered problem, are sometimes people that don't really think through what the needs and challenges are in rural communities or in marginalized communities or as women as opposed to men.
So we need in terms of the development community the ability to bring in the expertise that people have about their own needs and how to best meet them is an important point, I think.
>> ALLISON GILWALD: Thank you so much, Peggy. I think both of the questions relate to all of the panels and questions so perhaps we will move onto the next panel.
>> PABLO HINOJOSA: I think that will be really good, yes, I think we can go into the second question.
>> ALLISON GILWALD: And then when we get more questions from the floor, the first panel can also answer those. The second block of this discussion was really focusing on balancing innovation, access, inclusion and rights and very tied to the first one, and, of course, to the third one as well.
But here we have been a bit more concerned with the adoption of these new technologies increases, how should different stakeholders balance questions of innovation, access, inclusion? These are often being presented as tensions, regulation and innovation, and I think there is an increasing realization that under certain conditions you need to govern in order to create the conditions, equal conditions for innovation. And I think we only have to look at the concentration in markets around innovation that we currently see. So what are the rights implications of that, of enabling equal participation in innovation?
And some practical examples we hope will come up in the session. Anita, let's start with you, a lot of your work is focused on digital inequality and trying to surface the issues of economic justice and the right to development, which is built on very well from where Peggy ended off.
>> ANITA GURUMURTHY: Thank you very much. During the preparatory meetings we have had, Pablo has given me the bottom line. He has insisted we should make this a spirited discussion so that's what I'm going to do, and hopefully pick up from the very valuable comments. Broadly I think the balance today between innovation and inclusion is achieved through an approach that carves out a human rights-free zone for business.
So that some minimum threshold is met for the so called innovation economy to go unhindered. This is not incidental. We have research after research that points to how sustained lobbying by the big actors actually achieves this. So the balance is really not a balance that is from the sense that Allison pointed to, that we are all on the side of the people. The balance is not really coming from this. The first point I want to make is that technology foresight in the current environment becomes an exercise to enable ease of business with minimum negative externalities, so the idea is of negative externalities and the idea is not society. It's the economy first.
Just as an example, and this will be true for other regions as well, EU scholars have shown how the GDPR's balancing act between privacy and the data market has ended up leaving serious gaps in relation to algorithmic profiling and what it is doing to user autonomy which is trying to be fixed through the Digital Services Act, and the DSA is struggling to fill those gaps. So it's not as if the problem is unknown to us, we knew the costs of surveillant advertising in terms of what it does to autonomy, but we ignore the risks of the model. So the social harms and injustices are baked into the innovation. They are never fully acknowledged and as civil society, as concerned bureaucracies, in fact, competition commissions in many countries, as ethical technologists we have painstakingly put together evidence to mobilize opinion around the evidence, but policy ends up going from one fix to the other in some minimalist way without really transforming things.
So you are just tweaking things at the edges. So to do things differently, we need to envision the macroeconomic context very, very differently, and we need to answer the question, how can data rights bring new capabilities, new opportunities for social innovation?
And the second point which is a brief point I want to make is that the innovation today is built almost exclusively for market capture. So the incentive is market capture. This means business entities resort to aggressive preemptive patenting, the use and abuse trade secrets to lock up data, exploit loopholes and trade rules to avoid fiscal responsibility and algorithmic transparency. These are the several examples that we see in the world.
So the paradox is that we can harmonize data standards all we want across the globe like my colleague said, and we can push for interoperable data standards, but the open innovation culture our digital economy is built on unfortunately promotes cannibalization of smaller businesses and not the diversification of innovation. I will stop.
>> ALLISON GILWALD: Thank you so much, Anita. Alex, perhaps you can pick up the challenge there. Tell us a little bit about what efforts you have undertaken to promote human rights in that innovation context and, of course, in that open innovation culture.
>> ALEXANDRA WALDEN: Thanks for the question. Great to be included in this dialogue today. Where I will start is just to say I'm proud to be at Google where we are a company that is committed to Human Rights and that is not to say that we have got everything perfect, but it is, I think, an important step for any company to take as a baseline to say that you have a commitment to Human Rights, which really should be grounded in the UN guiding principles on business and Human Rights and if it's not grounded there, then I think the company has a long way to go.
So that's a baseline for how we think about these issues. And specifically when it comes to AI, you know, across the company we are really focused on and hopeful about the benefit that's AI can bring to everyone, individuals, consumers, communities, businesses, Governments, services, so we see that potential and we, that is what we are innovating to hopefully improve and bring technology to. Obviously there may be risks and we are focused on that, so internally our leadership talks about this to everyone who works at Google and they say this externally as well, we are focused on being bold and responsible when we innovate. So we are looking to push the envelope to develop new things to take on new challenges and in doing so making sure that we are having an approach, using standards to address the risks, and having Human Rights be an important part of how we think about that.
Another piece of this is really that there is a role for individual companies to think about how Human Rights play a part in their own work, and then how we work together as an industry on that. And then how we do so in concert with our partners at international organisations and civil society and with Governments as well. So there is a role for everyone in this.
For our own part, we are focused on ensuring that we have policies for how we do training, have policies for how we think about what outputs are allowable and preferable, policies for how our products are going to work. All of those things are ways that we need to take responsibility for the things we are developing or deploying through our products.
We also have frameworks for how we think about obviously how we integrate Human Rights across our product suite, and then we have a set of AI principles that are, again, how we think about our own work.
Obviously though we think that this work is so important and pervasive across the industry and globe that any individual company should be doing this, but is there is an important role for regulation that AI is too important not to regulate, but that that needs to be done well and smartly.
So in that way, we are engaging with civil society, international organisations and Government about how that should happen. Pre-existing standards that we already have from the UN guiding principles to work happening at the OECD, and various other places, those are standards that we should be using as a baseline for how we think about this, and to be informing regulation. So there is a lot of good stuff out there, and we really are, and everything is moving fast, but we are in sort of early days and I think ensuring that we are not reinventing the wheel, but taking the frameworks we already have and integrating those is the most important way that we can proceed.
But, again, I think creating a level playing field for the industry is really important and Human Rights must be the baseline, and I think a company that is not explicitly committed to the UN guiding principles cannot say that they are responsibly developing AI.
>> ALLISON GILWALD: Thank you so much, Alex, if I could do a follow-up question again right after that. Could you give us -- you have spoken about Google is very aware of their risks associated with this, Human Rights risks associated with this and that you have certain policies and training and frameworks in place. Could you give us a practical example?
What do you see as the biggest risk? How are you mitigating that with these training frameworks? If you could identify one? I know there are multiple. And then if you could just, you said regulation of the industry is important.
What do you see as the biggest issue that requires regulation?
>> ALEXANDRA WALDEN: One example I will share on the sort of risk side is obviously around things like deep fakes or synthetic content that is manipulative. That is something we are very concerned about. We know many Governments and members of civil society and others are concerned about, so for us we sort of do a few things.
One is we make sure that when we are thinking about how technology gets integrate need our products we are thinking about how to flag things for users so they can identify this type of content. So on YouTube we have content creators creating basically indicators on the content to demonstrate where something has had AI manipulation or the use of AI in the content to help people understand what it is that they are looking at.
There are obviously many challenges with that, but it's a way that we are thinking about how we can be helpful to people as they are engaging with content. Another example is through Google Cloud we created something called SynthID which is a way to watermark content to identify it's provenance so you know where it's comes from. That's another way we are thinking about how people can figure out what is authentic content or the provenance of content, where it comes from. That's what we have done on our own and what we are putting out in our own products. We do this work in concert with other colleagues from industry and civil society with the partnership on AI and C2PA where we are thinking about protocols around this.
So that's sort of one area where we see challenges related to AI, and where we are working in our own products and then certainly with others across industry and other sectors to hopefully improve and address the problem.
When it comes to regulation, there are a variety of things that we think about when we are thinking about what could be useful in terms of both preserving the ability to innovate while also making sure that we are maintaining standards related to. So in particular that is making sure that we have a sort of proportionate risk-based framework that's focused on what are based in actual likely harms so that we can have tailored regulation. We can always sort of improve and iterate later but broad big regulation at the outset is both harmful to innovation in this area, and really I think doesn't help us have a narrowly tailored solution to the problem.
Also differentiating between types of actors, regulation should be distinct between what are the responsibilities of the developer, a deployer, or an end user. We also sort of are very focused on things whether it's more of a hub and spoke model that there shouldn't be a single AI regulator, but that every agency over time is going to need to be a regulator of AI in some way. If you regulate the financial sector, you will be dealing with AI. If you regulate education, you will be dealing with AI. So really there needs to be capacity building among all types of regulators, not single, not a single focus.
And then maybe the last thing I will flag is about interoperability and ensuring that where individual countries are regulating, multi- nationals are operating everywhere, so we want to make sure that there is alignment and interoperability between the different regulations, and that they are taking advantage of the standards that exist, many of the ones that I already mentioned, and, again, using the UN guiding principles as the baseline for when you are talking Human Rights. Sometimes we see pick and choose concepts but then maybe muddle them a little bit and I think that is confusing for those of us at companies that have Human Rights expertise and for all of you are stakeholders who are Human Rights experts and trying to operationalize those things.
>> ALLISON GILWALD: Thank you so much, Alex. Julian, perhaps you can come in from that perspective. Alex has been speaking about some of the company and industry measures. And I think importantly also indicating what can be done at the production level, the kind of technology, by the design level but that doesn't necessarily deal with some of the either unintended consequences or outcomes, negative multipliers that Anita referred to.
What do you see more broadly as some of the challenges and particularly the work that you've been doing around Internet Governance as a kind of under pinning if you get that right for AI?
>> JULIAN THESEIRA: Thanks Allison for the questions and so on. Just, again, more broadly thinking about the scheme of innovation, Human Rights and so on, so and coming back, again, to the UNESCO recommendations, so something that CIDP strongly supports is as well the idea of sort of red lines against certain types of practices or AI systems that could be, could represent or could result in gross violations of human rights. Or fundamental rights and these are not just ideas from civil society. So, for example, UNESCO recommendations already recommends provisions of AI for social scoring, AI for mass surveillance, recommends a consideration of environmental impacts. AI also has a material dimension, so it requires resources of various kinds.
And these are recommendations that countries have already agreed to, so now like next step is to implement them and we see an example in the EU, there is another AI act that prohibits sort of high risk AI practices, and we encourage more countries to take the next step of implementation and ensure that certain high risk practices are prohibited.
And then linking back as well then to the question of Internet Governance more broadly, I think some of the issues related to Internet Governance are still very much relevant for AI governance and policy. So one of the building blocks of AI is data. So models need to be trained on data.
And we already have in many countries around the world, there are data protection authorities, there are data protection, various types of privacy frameworks, data protection frameworks so we need to ensure that the frameworks continue to be implemented and necessarily updated for counter advances in AI and emerging technologies.
>> ALLISON GILWALD: Thanks very much, Julian and be aware that we are going to take a round of online questions or maybe in the audience. Pablo, you will decide.
But before we do, Anita, can you just respond to some of the inputs that have been made and just to return to your original statement?
>> ANITA GURUMURTHY: Thank you very much. I just wanted to respond to my colleagues here. One is that I think tremendous work has happened from the different UN agencies, you know, the action line holders after the WSIS. But I do still think that more work needs to be done. One is that we somehow seem to have acquired a certain consensus around all Human Rights offline must be protected online.
We see this come up even in the WSIS+20 elements paper. But this at best is a partial view. It's not sufficient to deal with the complexities of network existence.
For instance, take digital ID programmes, deployment of facial recognition technologies in crime, social credit scoring for pensions. All of these examples show how a person who is redesigned by the matter in which tech renders who we are. Again, other example is about harms in data value chains, the livelihoods of those who may not be users.
So we need to account for those who are non-users. People who mine minerals, those whose lands are given away to data centers who are dispossessed. I wanted to give an example the research by the EDC of hedge funds and big food companies who operate in future markets because they control the data of farmlands. They have access to climate data. They are able to project the price of, say, wheat, in the global market, and they are able to manipulate the pricing and bet on the pricing.
And just take a small farmer who really unbeknownst to the small farmer they are actually having pricing changes in future markets because of AI. So you are actually talking about Human Rights not just of users, but of non-users. So we need to see these new afford answers and programmes for these rights and these are guarantees that we need in the social reorganization by technology. Which is why the ILO, for instance, is looking at upping its standards in decent work. We need to look at algorithmic management as an essential core of decent work standards.
The second point is about the product lifecycle. Again, made by my colleague on the panel. And I think that looking at the product lifecycle is not adequate because we need a wider society view of innovation, a kind of call for learning view which is that we need to look at AI, we need to look at the AI economy as always embedded. So this idea of an always embedded economy meaning an economy that is embedded in society and social choices and not the other way around.
We look at society is embedded in economy and try to tweak. So there is also in para 22 of the GDC a call to Private Sector which really is about voluntary responsibility. But I think in the global political economy of things if we want to move from data extractivism to equitable knowledge economies and pluralistic knowledge societies we may need other more compelling tools as a global community to bring greater accountability to private actors and soft exultations.
So lastly, I just wanted to say that the WSIS + 20 elements paper in para 78 talks about how many Developing Countries continue to face significant barriers in harnessing digital technologies due to limited technical expertise, weak institutional frameworks and constrained fiscal space.
For a long while now I have been very, very exasperated with this kind of characterization of Developing Countries. It's not as if these constraints are because we are inherently lacking. These constraints emerge because of the context, and they are not historical, they are not because of any inherent deficiencies in our societies. They are manufactured by an extractivist economics. Look at the debt, look at the debt burdens on about 45 countries.
And these, this kind of extractivist economics is historical and it's also baked into the digital paradigm. So the bottom line is whether there is international courage, not just political will to remove these constraints even as we enable capacity building, and I guess all Governments need capacity building.
So I would really want the elements paper to acknowledge that these are not constraints that are owing to these Developing Countries. These constraints are manufactured. So that's something that we really need to remember.
>> ALLISON GILWALD: Thank you, Anita. Let's go to the audience.
>> PABLO HINOJOSA: There are two microphones and I barely can see you so please stand up if you have questions.
>> AUDIENCE: Thank you. My name is Altina. I will be speaking in Spanish. My question is on the regulations and the tension against, between innovation and the respect for human rights. One thing that I found interesting is that companies have to have a special policy on human rights. I mean, that's strange. It's evident, isn't it? You have to respect them. That's just it.
Now, we have a parliamentarian here. I would like to know what is the vision? How do you see things in Latin America, for example?
>> AUDIENCE: My name is Christian Amigo I'm from the Democratic Republic of Congo, the DRC. My question related to the renewable tech relies on Cobalt, as you know, which is mined by children, with the labor of children. So how can the precautionary principle demand ethical supply chains before labeling tech as green.
The second one, how should the precautionary principle mandates ethical supply chain audits for AI developers before deploying models and how do we enforce this? Thank you very much.
>> PABLO HINOJOSA: Thank you. Malorie.
>> AUDIENCE: Malorie Nodle. I wanted to just reflect on my many years of experience in Internet Governance, particularly in standards bodies because what I have recently realized is the role of regulation in helping enforce some obviously use of Human Rights tools, but that standards also plays a role there. So if -- I would love to hear folks' reflection on what I feel is not so much emerging anymore, because for myself, I have been doing the work for 10 plus years, but definitely a strengthening of this idea that Internet Governance in particular in the development of technical specifications in an open multistakeholder way has found alignment with regulation that can really enforce the adoption of goods standards, hopefully in service of Human Rights.
So the work that you are all doing and anyone who has reflections on that, I think it would be really nice to hear. Thanks.
>> PABLO HINOJOSA: Thank you very much, Malorie. So Agustina, we will wait for your answer to come in the third block when Rodrigo will speak.
Christian, thank you. I would like to ask if someone from here would like to take that one. And Julian as well, of course. And Malorie, very important one because there is this element of protocols and discussions about let's think about it from the start, and I think that has been part of the conversation.
And we have Tim as well, thank you, because I know he is in Australia and following it at the very odd hour, so that's very nice. I will not be able to read all of what you have written, but please join online to see very important comments from Tim. Let's go ahead. Maybe, Anita, you can start.
>> ANITA GURUMURTHY: Thank you very much. I just wanted to address two points. The first one is about the precautionary principle. The people on the panel who may know more than me, but I would certainly like to agree with you and say that the larger implications of an intergenerational ethics also implicating planetary wellbeing should be programmed into the way we look at AI ecosystems. So we are looking at value chains today in a very, very limited manner. So you have due diligence rules for the EU, for OECD, but all of this really look at a very limited idea and understanding of harm. And I think that we should move beyond the product to understanding people, understanding intergenerational justice and understanding planetary boundaries.
The second one is to Malorie. And I think that it's indeed very important to, if I understand your comment correctly, I think it's indeed really important to bring progressive values from the technical community around openness and interoperability into regulation in an appropriate way, but I just want to caution here that we should neither overvaluize interoperability, nor demonize fragmentation because I'm just reminded now of President Trump's administration and the big beautiful Bill where every data set is sort of integrated and made interoperable. I would really caution against wanting to apply these very important technical values as my friend here said into our future regulation, but do so in a very context specific way, and contextual integrity, therefore, matters very much when we speak about technical values because there are also larger political ideals and values core constitutional principles we have to abide by. Thank you.
>> PABLO HINOJOSA: Peggy.
>> PEGGY HICKS: I want to add in the point of supply chains raised by the gentleman from the DRC. I absolutely value Anita's comments about the broader framing that needs to be here. I do want to emphasize that there are tools that currently exist that could have a significant impact if they were actually used in the way that they ought to be, and the reality is that we do need to create the right types of both incentives and disincentives for companies to actually do the risk assessment that needs to happen to ensure that those supply chains are not using exploitive labor in the way you have talked about.
And right now there are some movements towards mandatory Human Rights due diligence, the CSDDD and the EU for example, but there is also a movement to water that down some as well, and to make sure that we are only looking at sort of the supply chain side and the end users and not looking at the overall human rights impacts of the technology that's being used. In the broader sense Anita is talking about as well.
The reality is if we do leave it up and this is what Alex was saying too, if we leave it up to each of the individual companies and actors, we are creating an environment that is ripe for the type of exploitation that's been talked about because the incentives are for everybody to just deliver what they can, and some companies will commit in some ways, and be at a competitive disadvantage because they do it compared to others that don't.
So we need Governments to step in. And we do have the legislative tools to do it and the basis upon which to do it is the UN guiding principles that have been mentioned so we need more of that going forward, I think.
>> PABLO HINOJOSA: Pierre, could you answer the question.
>> PIERRE BONIS: Yes, to pick up on the integration within the standard development which are issues linked to human rights, on that part I might be a little bit provoking, but in the AFRINIC we are in the technical community and we do not like technocracy.
For us technical community it is not up to us to determine how to best protect Human Rights and standards. It's not our work. It's not our role. And we do not have this competence.
It's not up to us. We have seen several of these examples with the UTF, we have seen several works and on the RFC, the PCF, the press for comments. And we have seen biases that were absolutely unexplainable and people had goodwill indeed. They said the more intermediaries we have the more dangerous it is and nobody knows why having many men is dangerous. I don't know why somebody said that, maybe somebody wrote it and maybe somebody wrote it at a time where they were having problems with middle men, and at the end it's the standard. So I think that's the dangerous path.
I'm not saying that technology is so neutral that it should not address fundamental aspects of human rights, but imagine for a moment, imagine that we could guarantee these rights through standards and protocols. I think it's wishful thinking, and if you avoid difficult conversations, well, difficult conversations are the assessments of technologies, what kind of harmful impact they can have, how can we go back in time and try to limit these harmful effects? So we can't by design from the start solve all problems. It's a techno solution that does not work. We have seen it in all history. It never has worked.
>> PABLO HINOJOSA: I'm going to speak in Spanish. I think we should move forward. And we have Alstina's question which is to do with the regulatory part and the policy issues and I think in that sense it would be appropriate to give the floor to Rodrigo I would like to hear your view on the legislative work. Its preliminary or is it post facto? So what comes first, the chicken or the egg? Which way should it go?
>> RODRIGO GONI: Well, that's good because if we are the last technical revolution, wait until after, they had all of the problems on the table, and the problems had consolidated, that's when the Parliament did. Publicly the Parliaments would wait and see the behavior, the problems well defined that's only then where the Parliaments would look at the problems and try to correct them.
But what happens now in the current context is that there is a permanent change, constant disruption. We can't behave that way anymore. We can't attack the problems that way. If the Parliaments want to have a role as a player, proactive role in the defense of Human Rights going forward, you have to face it, you have the protection which is a duty we cannot avoid as a parliamentarian it is our duty to see it the way it is. Google doesn't care. Google's function is just to develop technologies. We all have our different fields.
The role of Parliament is that they cannot escape is to protect the Human Rights. Human Rights means protect people against harm, but also to facilitate that they exercise. We talk about access, for example, we talked about companies being able to develop. Well, that's where the role of Parliaments is not only legislative, but it also handles the resources of the country.
So if we really want to expand the technological capacity so that everyone really can develop this technology, then that's a push for many Parliaments to act.
I would like to say the following. We can also be substituted. We have discussions in the Parliaments that are quite heated. We are also threatened by being substituted. We would have AI would be much better than us. We are MPs, we discuss full day, whole day, whole weeks. AI would do that much more quickly, but where I'm heading is if we as Parliaments really want to fulfill this basic role that we have, fundamental role, we have to change the paradigm.
The paradigm for Parliaments will have to change towards being proactive from being reactive. It doesn't make any sense to run after technology and change it. We will always arrive late, and in the wrong door. Because we want to hurry up, we will just forget Human Rights at the other side of the door.
But we have to change the paradigm as the following way, for example. We have all discussed all day and yesterday the local against the global rights and laws as opposites but to have a correct approach, we have to move beyond that. Another thing that we have to move beyond, for example, is this multistakeholder model. We see its function, we all agree, in Internet Governance. Now, if the Parliament want to arrange things or play a part, they have to enter into this multistakeholder perspective.
We talk about legitimate citizen representation in the Parliament, yes, but in this context we would have no other possibility but to sit down. We are doing that in some Parliaments, sitting down on an equal level with industry, Academia, and civil society. Why do we do that?
The level playing field, it's not because we are convinced. It's because we have to. Legislators have no other possibilities, at least not most of us. We can't follow really the events, the technological development. It's impossible for us to be on top of it. Because you are 24 hours, and then you will not be reelected. You don't have enough hours. So you have to change the paradigm.
You have to do this in in the Parliaments quickly because we all agree that we need regulations except maybe some, we know what voices we are talking about that confirm the rule of the exception. Everybody else agrees that yes need not local regulations. We are quite clear we need global regulations that will have to take into consideration multistakeholder perspective, and it will have to involve local Parliaments because citizens will not allow Parliaments to stay out of this. I'm about to finish, Pablo.
We have to do this with an anticipatory focus. We have to be flexible. We have to have a focus that allows us, we have to try to establish a framework where we can develop these dialogues.
We have to have a sand box, a regulatory sand box which is much more, much wider, much bigger than we have had so far. So we can test out things on a much bigger scale.
We have to do this, Parliaments, otherwise human rights will be breached not only the second generation rights, but also the first generation and also the risks are very grave. If we don't do it now in the way I have described, probably we will arrive late at the party and problems will be of such a magnitude that we will have this very brisk shifts that are much more harmful.
>> PABLO HINOJOSA: Thank you very much, we are in the need to start wrapping up. There is a lot of reframing here. I'm so sorry, one second, please, I hope I can give you a chance, but we need to start wrapping up. Can you do it in ten seconds? Okay. Let's do it.
>> AUDIENCE: Hello. I would like to make a contribution about the neutrality of the Internet because I see it in a different way. In Africa the question of the Internet neutrality is a different one. I come from Senegal and there is, there was a pre-electoral campaign. There was a lot of Internet campaigning, and where is neutrality in this case when we did not always have access to the Internet?
>> PABLO HINOJOSA: Yes, and I also have Christian's question in the air. I think we need to start wrapping up. I would like to start with Alex followed by Julian, followed by Anita, and hopefully we can have a chance for the others to just complement what are the key takeaways of the panel. Alex. I'm putting you on the spot with this.
>> ALEXANDRA WALDEN: You are. The key takeaways. I think for me it reinforces all of the ways in which we at companies need to be engaging with everyone in civil society, international orgs, legislators at the national and local level to ensure that we are getting this right and not that it's a tension, but there is a lot of work to be done to figure out just how we get the balance right.
>> PABLO HINOJOSA: Through, Alex. Fantastic, Julian, would you like to address some of the questions? Think about Christian and our last person from Senegal.
>> JULIAN THESEIRA: Thank you. So in terms of the questions from the floor and the last question as well, so here now I will speak in my personal capacity not as CIDP. So personally, I'm sympathetic to what is raised around the broader systematic concerns. I do think we need to think about them. I don't think questions about digital rights, Internet rights, AI governance can be divorced from let's say product concerns around justices or inequalities, inequities in the economic system, the global economic system resulting in things in inequitable Internet access in countries like Senegal, and there are ongoing conversations this year around debt forgiveness, debt justice. I think these are things that are worthying about, the international community should think about, and also there are new challenges emerging as well, such that emerging country debts are increasingly being held by private stakeholders whether they are traditionally other countries or multilateral development banks.
So I think these are, so I think one take away I would say is policy is interconnected with other policy domains and I think systematic concerns and we need to think about that as well. Thank you.
>> PABLO HINOJOSA: Through, Julian. Just to add into the complexity. Lagfort is on line and she is saying international cooperation has been drastically reduced and that is another part we knee to understand along with significant cutting in funding and grants that have been traditionally provided. So that is another situation that affects in particular the Global South. That's Elayne Ford. So if you can thread that into your comments, Anita.
>> ANITA GURUMURTHY: Thank you very much. I wanted to make four points very quickly. The first is that we know one thing very clearly now that we don't need models to be universal. We know that universal models can be totalizing and detrimental to diversity, so the application of AI can enhance development autonomy if local communities are put at the centre and currently the local is hollowed out and local value is transferred to the circuits of global finance capital economically and culturally. So there are four things for public interest governance. The first take away is that public interest governance is not just about harms mitigation. It's also about responding positively to societal needs, the new paradigm so you spoke about for intergenerational justice.
Secondly, public interest governance must respect what Pablo just mentioned, the spirit of international solidarity, extending accepted values that some of my friends in the audience spoke about, barring from environmental laws just the precautionary principle but there are other environmental law principles. The concept of polluter base. If you call misinformation, you better pay. Ideas such as common by differentiated responsibilities. That comes from environmental law.
The third point is that we need public interest governance in AI to recognize as I said the technical norms cannot become automatic standards for political norms. Take openness, ecological activists have spoken about how open genomic information databases have just become sources of bio piracy. So you really need that. And my last point would, therefore, be that we should really not have antagonisms between public governance and common spaced people alternatives. We need public infrastructure and we need common space alternatives from the other people.
>> PABLO HINOJOSA: I see a red mark saying time up so I understand that this finishes at 5:00. Allison, I would leave you to wrap up this. We covered a lot of ground, I think. If we didn't resolve all of the questions that were asked at the beginning, I think we left with more questions and that's a good thing.
>> ALLISON GILWALD: Thank you, Pablo. I think that's a big ask. I'm certainly not going to try and summarize everything that's said. I think what's been interesting in terms of developing some of the debates that have been on have come from everybody actually in different ways. But I think a number of the critical ones around the precautionary principle which I think has also been evident in technology debates over many, many years.
How soon do you step in to mitigate those harms without preventing the sort of innovation that might happen there? But I think this conversation even five years ago at an IGF would have been still pushing very much prevention of regulation in order to ensure innovation. And I think things have moved so fast and the potential harms and growing inequalities have become so great that I think just from the different stakeholders on this panel is an acceptance that something has to be done, that the people have to be heard and have to be protected from the Parliaments that we need to be looking at beyond just the very siloed digital rights and digital harms that we need to be looking at this whole value chain that goes from the extractive mineral base through to extractive labor, exploitive labor base, so the whole labor ILO work that's been done on decent work and detraumatizing work that's under way at the moment through to the extractive data debates we have been more traditionally having.
But I think the important points too that were made about enabling equal participation. So it's not only about harms mitigation. It's about ensuring access to what is actually public data that might be proprietarily held that protects individual rights and the harms that can arise from open systems.
I think there is an acknowledgment that self-regulatory models that we have had up to a certain point have failed. I think the band of European legislation has moved us towards that. But I think what is also very interesting is the acceptance that there is so much we can do nationally and regionally, but ultimately because these are highly concentrated big global companies that are dominant in this area that it's going to require some levels, high levels of regional, global cooperation if we are going to resolve this.
And that that needs to be done in a way that really reflects voices and vision and views from the Global South and from different multistakeholders, because some of that work has been happening at the global level, but has tended to be either multilateral high level Government without multistakeholder participation and then the multistakeholder participation has been in these more non-decision making forums. So I think connecting that came up strongly in the different views in the panel.
There was so much more and such interesting comments, but I think we have to leave it there.
>> PABLO HINOJOSA: It has been an absolute pleasure. Thank you for those that endured the 90 minutes. We will be online for the rest of the next decade or so. Thank you.
(Applause).
