Digital Policy Perspectives

30 May 2024 14:00h - 14:45h

Table of contents

Disclaimer: This is not an official record of the session. The DiploAI system automatically generates these resources from the audiovisual recording. Resources are presented in their original format, as provided by the AI (e.g. including any spelling mistakes). The accuracy of these resources cannot be guaranteed.

Full session report

Experts convene to advance equity and inclusion in the digital sphere at Broadband Commission session

During a Broadband Commission session focused on “Digital Policies to Advance Equity and Inclusion,” moderated by Leona Verdadero of UNESCO’s Digital Policies and Digital Transformation Team, a diverse panel of experts discussed the importance of inclusive digital policies and the progress in addressing the digital gender divide.

Selena Abdullah, the Broadroad Commission’s Executive Director, opened the session by addressing the issue of outdated digital footprints, using her own photograph as an example. She highlighted the narrowing digital gender divide, while acknowledging that significant disparities remain, with 244 million more men than women using the internet regularly. Abdullah emphasized the need for inclusive digital policies that consider various aspects of inclusion beyond gender, such as ethnicity, language, and disabilities. She also outlined the Broadband Commission’s initiatives, including the Equals Girls in ICT initiative and the partnership with GSMA to address the digital gender divide.

Stephen Lang, from the U.S. Department of State, discussed the U.S. International Cyberspace and Digital Policy Strategy, which promotes digital solidarity and an inclusive, secure, and resilient digital ecosystem. He detailed the U.S. government’s domestic and international efforts, including an executive order on AI safety and security, broadband deployment programs targeting underserved communities, and the Digital Connectivity and Cybersecurity Partnership Program to enhance connectivity overseas. Lang also mentioned initiatives led by Vice President Harris to advance women’s economic security and close the gender digital divide, with substantial financial commitments to support these goals.

Rumaan Choudhury, CEO of Humane Intelligence, addressed the need to incorporate diverse perspectives in AI design and policymaking. She highlighted the biases in AI systems due to training on data that predominantly reflects the global north and male perspectives. Choudhury described her nonprofit’s work in structured public feedback, expert red teaming, and bias bounty programs to evaluate AI models and improve their fairness. She advocated for a global community of practice to assess algorithms and prepare a diverse workforce for this task.

Anita Gurumurthy, Executive Director of IT for Change, presented a feminist framework for promoting digital justice, emphasizing that inclusion must be empowering and reduce inequality. She discussed the adverse inclusion and the connectivity paradox, where increased connectivity does not necessarily lead to benefits for all. Gurumurthy called for public policy that redistributes digital benefits, ensuring that technological innovation correlates with productive forces at all levels.

Mela Thiel-Gruensteg from GSMA discussed the latest mobile gender gap report, which shows progress in women’s mobile internet adoption in low and middle-income countries but also underscores the persistent challenges. She outlined the barriers women face, such as handset affordability and digital skills, and the need for targeted interventions to ensure diverse and regular internet use. Thiel-Gruensteg highlighted GSMA’s initiatives, including the Connected Women Commitment Initiative and training for policymakers on digital inclusion.

An audience member raised the issue of youth representation, particularly young girls from the global majority, in policymaking. Panellists acknowledged the importance of youth voices and the need for targeted approaches that consider the barriers specific to young women.

Cédric Wachholz of UNESCO concluded the session by summarizing the discussions and reiterating UNESCO’s commitment to a human rights-based approach to digital inclusion. He emphasized the importance of data governance and mentioned the Broadband Commission’s upcoming working group in this domain.

The session highlighted the multi-stakeholder approach to addressing digital inclusion, the recognition of the intersectionality of inclusion factors, and the proactive strategies proposed to prepare individuals for the digital world. It underscored the importance of considering the full user journey, from access to diverse and regular use of the internet, and the need for collaborative efforts to create an inclusive digital future.

Session transcript

Leona Verdadero:
Broadband Commission Session on Digital Policies to Advance Equity and Inclusion. My name is Leona Verdadero, and I’m with the UNESCO Digital Policies and Digital Transformation Team. I’m very happy to be moderating this excellent panel with distinguished leaders and experts on this very important topic. Let me begin by calling Selena Abdullah, Broadband Commission Executive Director, Chief of the Strategic Planning and Membership Department at ITU, to give the opening remarks. Thank you.

Sulyna Abdullah:
Thank you, Leona. First of all, I’d like to apologize for the misrepresentation of my photograph on screen. That is at least 15 years old. This is the problem when you can’t get photographs off the internet. But this is how I look like today. Hello, very pleased to meet you. So good afternoon. And thank you for that introduction and for the invitation to the Secretariat. I am the Executive Director of the Broadband Commission. And as Leona said, I also have a role in the ITU. Now, many of you here today are familiar with WSIS and the ITU, and the ITU-UNESCO Broadband Commission. But for colleagues who are here who are not familiar with us, these institutions share much common ground, extending the benefits of universal connectivity, improving digital policies, and giving a voice online to marginalized groups and communities. Now, I dare say that that is why the Broadband Commission has organized this session today about digital policies to advance equity and inclusion. Let’s take a step back and reflect. Why does digital policy need to be inclusive and equitable? Now, there’s a Chinese saying that women hold up half the sky. Today, arguably, women also need to hold up half the sky. hold up half the internet. And ITU was the first agency to measure and also quantify and monitor the digital gender divide. And we are seeing some progress. The digital gender divide is narrowing in most regions of the world, although last year, there was still 244 million more men than women using the internet regularly. And the statistics show that there is a 5 percentage point difference right now between men and women using the internet. But gender is only one aspect of inclusion. I’ve heard a lot here this week about the need for digital technologies to be inclusive and representative and to consider the perspectives and needs of people from different ethnicities, different languages, different scripts, or those with special visual or audio needs, for example. We are stronger when we are more diverse. And we can see, hear from, learn, and understand different points of view. Digital technologies and more recently, AI technologies and AI systems should be inclusive by design. For instance, we are seeing concerted efforts to develop generative AI for languages other than English, including Spanish, Arabic, and Chinese. But if we are already worried that a less than multilingual internet would promote content in widely spoken languages more than other languages, how will generative AI affect this? Will AI accelerate the demise of local dialects or help preserve cultural and linguistic heritage? And to achieve real inclusion, we need multi-stakeholder cooperation in action with many voices and perspectives reflected and taken into account. Both ITU and the Broadband Commission champion efforts to bridge the gender digital divide through working group reports and programs like Equals Girls in ICT. cyber mentorship. We are working for ICTs to be available and affordable and accessible, which means they are designed to meet the needs and abilities of as many people as possible, including those with disabilities. The Broadband Commission is made up of many commissioners who are actively campaigning for greater inclusion. ITU, for example, and the GSMA established the Equals Partnership to work specifically on the digital gender divide. And over the years, we’ve published different analyses and studies, for example, with Microsoft into the specific needs of people with disabilities for digital technologies. So we have achieved a lot. And we all recognize the vital importance of this issue. But much more, as we all acknowledge, remains to be done. Although the technologies and therefore the issues are evolving very quickly, the need to include and embrace many diverse voices in digital policies remains constant. And that is what this session is about today. Thank you. And I look forward to hearing from today’s speakers.

Leona Verdadero:
Thank you, Selina. Now I’d like to introduce our panelists. We have Dr. Rumaan Choudhury, CEO of Humane Intelligence, Broadband Commissioner, and US Science Envoy for Artificial Intelligence. We also have Ambassador Steve Lang, Coordinator for International Communications and Information Policy, Bureau of Cyberspace and Digital Policy, US Department of State. Anita Gurumurthy, Executive Director, IT for Change. And last but not least, we have Mela Thiel-Gruensteg, Senior Director for Digital Inclusion at GSMA. Just to introduce the format, so panelists will get a first round of questions, and you will each have four minutes to answer. Thank you. The first question is for Ambassador Lange. The U.S. recently released its International Cyberspace and Digital Policy Strategy, which emphasizes a rights-respecting digital future and highlights the importance of digital inclusion and gender equality. How does this strategy address the unique challenges faced by women in marginalized communities, especially in the Global South, and what specific initiatives are being undertaken to promote digital inclusion and human rights in this region? Thank you.

Stephan Lang:
Great. Thank you very much. I guess. All right. Is that good? Yeah. Okay. Thank you. And thank you for the opportunity to be here today. And thank you for the question. The United States takes very much to heart these issues that we’re discussing here today. And we’ve really tried to make equity and inclusion a central part of our digital strategies and policies at home and abroad, including the International Cyberspace and Digital Policy Strategy, which was launched by Secretary Blinken at the RSA conference in San Francisco on May 6th. And this strategy really aims to set the United States on a path to implementing an affirmative and proactive vision for building digital solidarity, connecting people and information to foster a more inclusive, secure, prosperous, and equitable world. And digital solidarity recognizes that governments need to implement their own domestic policies, but it seeks to develop shared mechanisms that will help maintain open, interoperable, and reliable digital ecosystem, as well as trusted cross-border data flows. And it works to foster democratic values based on rights-respecting policies and an inclusive multi-stakeholder ecosystem. Digital Solidarity recognizes that all who use digital technologies in a rights-respecting manner are more secure, more resilient, self-determining, and prosperous when we work together to shape the international environment and innovate at the technological edge. And in this regard, our work will really focus on three areas, or I’m sorry, four key areas. Infrastructure, governance, norms, and capacity building, all oriented towards digital solidarity for an open, inclusive, secure, and resilient digital ecosystem for all around the world, enabling new solutions to global challenges. So looking at how we’re putting this into practice, I’d first like to make a couple of comments about what we’re doing domestically and then focus on what we’re doing internationally. So domestically, the Biden administration in October of last year issued an executive order on the safe, secure, and trustworthy development and use of artificial intelligence. And this order establishes a process to develop new standards for AI safety and security and seeks to protect citizens’ privacy, to promote innovation and competition, and to advance equity and human rights. We can also look at our own domestic broadband deployment efforts under the National Telecommunications and Information Administration’s Broadband Equity Access and Deployment Program, which recognizes that digital access alone does not bring transformation or inclusion. Its digital equity programs, which are providing $2.75 billion, will help ensure that all people in communities have the skills, the technology, and the capacity needed to reap the full benefits of our digital economy. And it targets specifically historically underserved communities and tribal populations. And then on the question of how we can be responsive to the needs of the global South, digital and cyber capacity development are important parts of our digital solidarity initiative. The U.S. government has robust assistance programs devoted to closing digital divides, including the gender digital divide, and enhancing connectivity overseas. Through technical assistance, capacity building, and training grants, the United States Digital Connectivity and Cybersecurity Partnership Program, or DCCP, has helped provide partners the expertise and training that they need to develop and govern secure, rights-respecting digital ecosystems, and to leverage digital technologies to achieve the Sustainable Development Goals. And for the USG, these efforts are really government-wide. Last year, Vice President Harris announced in Africa several initiatives advancing women’s economic security and closing the gender digital divide, including the Women in Digital Economy Fund, or YDEF, which just last week – well, her office announced just last week an additional $500,000 in direct USAID contributions to the fund to further strengthen the digital inclusion of women and girls with disabilities, and $46 million in aligned U.S. commitments to USAID’s Women in Digital Economy Initiative. The Vice President also announced $145 million in new aligned commitments to this initiative from government, multilateral, private sector, philanthropic, and civil society partners to accelerate progress towards the G20 leaders’ commitment to halve the gender digital divide by 2030. And the fund and related Women in Digital Economy Initiative have now generated over $1 billion in public and private commitments to accelerate gender digital equality. The initiative is managed by an industry-leading consortium that includes CARE, the Global Digital Inclusion Partnership, and the GSMA Foundation. In addition, through the Equals Hub that’s hosted by ITU, Equals – Equals with USAID and the Gates Foundation will support the Women in Digital Economy Community of Practice as a distinct project of the Women in Digital Economy Initiative and a deliverable of Equals. And I’m also very happy to announce, to say that the U.S. State Department has awarded $300,000 in funding to the ITU under Equals to help reduce the many barriers faced by women and girls in harnessing the benefits of digital access, including in Burundi, Libya and the Dominican Republic. And this is really just a snapshot. There are other programs that we’re engaged in. Before I close, I’m hopeful that there’s one additional message that is clear, and that’s really the need for multi-stakeholder engagement, partnership and collaboration on these efforts, because no one country or organization can do this all alone. And this is a really pivotal year at the U.N. on technology issues as we negotiate the Global Digital Compact and prepare for WSIS Plus 20. We continue to articulate global action on AI technologies, and they need to be underpinned by principles of equity and inclusion. And we need to work to ensure that the benefits of digital technologies are widely available for all while still mitigating the risks. So we look forward to working with all of you to try to make that vision a reality. Thank you.

Leona Verdadero:
Thank you. And the next question is for Ruman on AI inclusion. Based on your work at the intersection of AI, human rights and leadership and algorithmic evaluations, how can we effectively incorporate diverse perspectives and stakeholders in AI design and policymaking to ensure that these technologies do benefit all communities equitably? Thank you.

Rumman Chowdhury:
Thank you. And thank you for your presence. And thank you for the for the question. This new wave of AI systems is quite different from the machine learning and AI systems of the past. intended to be general purpose. In other words, traditionally we designed AI systems with an ultimate goal. It could be to be a math tutor for a child or to identify cancer from an x-ray. General purpose AI systems don’t have a specific objective or goal. They are meant to interact with us in a quote, natural human fashion, trained on the data of the internet. So there’s a few things to think through given that framing. First is, you know, companies will tell you that, you know, the data it’s trained on is quote, the data of the world. And that’s wildly untrue. As we’ve talked about, there is a gender digital divide. There is a global digital divide. The information that exists that these models are trained on are primarily reflective of the internet that exists today, which is primarily male and primarily global north. And thinking through that, as we think through the potential for bias and discrimination that can exist embedded into these AI models, you’ll be unsurprised to find that when you have a dataset that is largely driven by one specific population, the output of the models may not be so friendly to other populations. Now we’re seeing an increasing number of specified use cases that can be quite impactful to a broad range of individuals. For example, a few weeks ago, journalist Leon Yin at Bloomberg did an analysis of resume bias in hiring for ChatGPT, designing code to demonstrate how thousands of algorithms that were for men or for women would get different feedback and input if someone were to create some sort of a resume, you know, a parser or a resume evaluator using ChatGPT. Now we can imagine that this bias also exists in other forms of media as well. So image generation, the news outlet, Rest of World did a really stunning analysis that showed that when you typed in phrases like Indian man or Arabic woman, it resorted to these very traditional and sometimes offensive stereotypes. Frankly, personally, most of the Indian men I know are just wear Western clothing or suits. And yet the images that overwhelmingly came up were of sort of a very traditional Swami looking guy with a turban and a dot on his head looking eight old and wizened. And yet, by the way, if you type in American man, you will get a wide range of different perspectives of what American men may look like in different settings and situations. One that I will highlight to you that I think every time I give an example for, and then there’s gasps in the room, in that same exercise, they asked AI to generate an African doctor. The AI model would only generate images of a white man treating poor black children. It did not at any stage generate a black man who was a doctor. So just to give an example of how pervasive the biases are in the system, because again, it’s trained on the data of the internet, which is not actually a data of the world. So what can we do about it? My nonprofit Humane Intelligence is engaging in a global community of practice to create new methodologies for what I call structured public feedback. This is ways for members of the public and groups are not traditionally engaged in Silicon Valley to be part of directly being a pipeline of information from their thoughts, feelings, and expectations and experiences working with these models directly to the model owners. So there are three ways in which I do this. First is through expert red teaming. Expert red teaming is generally with individuals who are experts in a particular field of study and they’re brought in to identify issues in AI models. Public red teaming is the second one. The public red teaming is something new and quite different. Public red teaming is actually bringing in people with their expertise is actually their lived experience. This could be speaking a non-majority language, just could be being of a generally underrepresented. group online or in Silicon Valley. Last year we did the largest ever generative AI red teaming exercise. We actually did this hand in hand with companies as well as the White House and civil society. We had 2,200 people come in and evaluate AI models against the United States blueprint for the AI Bill of Rights. So what was interesting and revolutionary about this was this was more than just evaluating for cybersecurity. We looked at information integrity, bias and discrimination, inconsistent information, falsification, political misinformation. And we had, we learned quite a bit. The, and the last program that we’re working on our Bias Funding Program. So these are online competitions that are meant to build a skillset and educate people on how to interrogate and analyze algorithms. The last thing that is critically important is building out the community of practice of people who will be working on assessing algorithms. Now we’re seeing laws being passed all over the world. In the EU, we have the EU AI Act in the United States. There are multiple laws at the state level, some being proposed at the federal level. But I will say from my experience in industry and also in doing this work myself, there’s an insufficient number of people trained to do this work. Now, whether it’s in the US or anywhere else, there was no reason why this should be limited to people of a certain geography. Bias Funding competitions are open to anybody in the world. They are cash prizes. They provide challenges based on beginner, intermediate and advanced, and anybody can engage with them and learn from them. The results actually sit on our website. They’re all open source, open access. So if you imagine the kind of educational advantage that Kaggle gave to people all around the world, right? The ability to interact and access code on how to do data science. The goal is that we do the same. So all that is to say there are approaches that companies are using that people all over the world can be engaged in participating in to give direct. feedback for the models to be improved. And the second part is a challenge and a question for those all of us in the room is how can we better prepare a future global workforce, not only in the global North, but especially in the global majority to be part of evaluating, not just building AI systems. Thank you.

Leona Verdadero:
Thank you, Ruman. And on a very related topic for Anita, so you work on research and advocacy on data and AI governance and also feminist frameworks on digital justice. Could you talk to us more about what it means to apply a feminist framework to promote digital justice? Thank you.

Anita Gurumurthy :
Thank you. Honored to be here on this panel. I just wanted to clarify that feminist frameworks certainly pertain to social hierarchies of gender, but not only. And therefore it’s important to think about this lens as something that addresses all kinds of social differences and social power. And the most important lesson in the field for us is that inclusion into digital society and economy, and this is something that has emerged in the years after the WSIS, inclusion could be adverse inclusion. And what is adverse inclusion? Well, you are seemingly part of the benefits ecosystem, but that comes at punishingly high costs. Take the case of having access to social media, but being victimized by trolls, or being part of national ID systems or health ID programs only to be exploited by unregulated fintech companies trying to sell you insurance. Or if you’re a small farmer in a cooperative, you see that you’re contributing to common data spaces or data exchanges and you’re receiving nothing, but agri-tech companies are happily using your data for profits. So if we want meaningful inclusion, then we need to think of what can enable greater bargaining power for those that are less powerful in society. We are not just talking about opportunities here. but a shift in social power. That is why inclusion is meaningful only when it’s empowering. Today we are experiencing what is called the connectivity paradox. Even though more people are connected to the internet, only those who have the resources to make productive use of the internet are genuinely benefiting. This means digital inequality is a complex problem and meaningful connectivity is important, but not enough. We need to tackle other antecedents of social and economic power. A great example of this is the universal payments interface in India, a revolutionary digital public good. And while you see the QR code sitting on top of all the vegetable and fruit carts, the street vendors are not getting richer. A holistic society-wide approach to digital inclusion therefore has to be multi-layered. It’s not just enough to use a digital public good or have connectivity. Technological innovation needs to be correlated with productive forces at global, national, and local levels. And today, while some nations and some interests within nations are able to harness the benefits of these innovations, others are not. The transformative potential of information societies is yet to contribute to human freedom and development. And the first principle therefore in the feminist playbook is that we need to be sure that our interventions can contribute to a reduction in inequality. The data and AI regime has led to adverse inclusion that’s really worrisome. You may not know it because you’re a peasant or a fisher person, but you’re already datafied. The individualism of gadget centricity of tech diffusion has access to benefits of data, has converted access to benefits into access to products. It’s a product mode. Not only are we passive consumers, but we are also, as is infamously said often, the product ourselves. Indigenous people, peasants, cities, many collectives have challenged the idea that data is solely. individual. Indeed, some facets of data are individual and even private, but the aggregate collective data that holds societal insight is collective world and it’s part of social commons. So, the second principle here is that empowering inclusion enables local knowledge to be leveraged for collective well-being. Now, I move to the third insight. There’s also a need to look at the digital as a connectivity policy and a platform and data policy as vital to public law, not just private contracts. There’s a quality of the digital that’s paradigmatic or life-changing. As they say, the digital reorders society. It’s the new infrastructure of social organization, which is why public policy for affirmative action is vital. In India, the telecom regulatory authority ruled that net neutrality is important because the internet is an experience good. So, there can’t be one internet for the poor and another internet for the rich. Similarly, there are policies at the sub-national level in my country where the state is providing a cloud backbone, a platform backbone, and there are innumerable data trust communities that are benefiting from this. Therefore, the third principle from the feminist playbook is that the state is not just a provider of welfare. It is a catalyst for digital transformation. It’s an allocator. It distributes, it redistributes, and this is really key. And finally, about data and AI. We’ve been long arguing how algorithms must have diversity and representation. Representational diversity, however, is not enough. Even if algorithms have diversity of input, we may not see this result in representational justice because society is really not quite fair. And the ecosystem in which algorithms operate need a social approach, not about adding and stirring inclusion, but one that can achieve epistemic justice. We don’t want to see recidivism algorithm push more younger people. from marginalized communities into incarceration. We need to use AI as a means to take decisions that break traditional social hierarchies. So I want to conclude with the fourth feminist principle here, which is about a radical politics of inclusion, where the structural infrastructural power of AI is deployed to change the power equations in the social ecosystem. Thank you very much.

Leona Verdadero:
Thank you, Anita. And for Mila, from GSMA, we know that GSMA just launched your latest mobile gender gap report. So can you talk to us about the progress and women’s rate of mobile internet adoption over the last year, specifically in low and middle income countries? And how is GSMA working to narrow the gender gap in this regard? Thank you.

Melle Tiel Groenestege:
Thank you very much. And thank you for the kind invitation of the Broadband Commission to have the GSMA on this panel. So there is finally a glimmer of hope in terms of progress on the gender digital divide in low and middle income countries. We’ve seen progress year over year from a gender divide of 19% to 15%. But that means we’re now back at the levels we had before the pandemic. So there’s some optimism, but also let’s be realistic about the significant challenge that is still there. As you mentioned, there’s more men online than women. And that’s about the size of Indonesia, the fourth most largest country in the world. I think what is critical as we talk about the gender divide and all these statistics that we hear at these conferences is that we realize that behind these numbers are real life’s hopes and dreams of women and people in general. So we need to focus our efforts on addressing that issue, especially as a community, as we gather here together with quite a large impact on the digital futures of people. We cannot be complacent and maybe we should actually be quite upset about this persistent gender divide. So how can we address it? It’s really looking at the situation that women and digitally excluded face in the different countries. So that’s something we research in a mobile gender gap report. We look at the barriers that women face of accessing mobile internet and we find this is quite consistent year over year that there’s a number of barriers they face. Most important would be the handset affordability barrier. So women often don’t have the income or ability to purchase the device which has a high upfront cost. There’s also a lack of digital skills that is the second most important challenge. And then other issues related to is the internet relevant to people in that part of their life. Is there safety at security risk or other access barriers related to do I have electricity to charge my device? Can I actually go to a retail shop? Who is the sales agent in that shop? Is that a man who I don’t feel comfortable with entering into a transaction? So these are the barriers for mobile internet adoption and that’s also the numbers that we just mentioned in terms of the gender divide. It’s about adoption. But actually what does adoption mean? That’s someone who uses the internet once in the last three months. So maybe that we need to look at regular and diverse use of the internet. How do people actually go online and is that once a day and is that for more than just two applications or services? Is that a rich use of the internet? And then we actually see a very different barriers emerge. The biggest challenge there is safety and security concerns followed by issues related to data affordability but also quality of network experience. So we see that we need to take quite a targeted and different approach as we look at the user journey from adoption to diverse and regular use. And the GSMA is working on several levels to tackle those issues. So the digital inclusion team at the GSMA supported by GSMA members, but also funded by donor countries and foundations. We work with our private sector members, but also we support policymakers and we provide the critical data and insights to inform action. So one of the initiatives that our members have signed up to is the Connected Women Commitment Initiative. Twenty-two partners have renewed their commitment to increase their women customer base in mobile or financial services. And since 2016, since we launched that initiative, we have impacted the lives of, together with the partners, impacted the lives of over 60 million women in low and middle-income countries. So it really shows that when there is the commitment, when there is the platform for action, when there is the information and the support, we can achieve real change. We also trained 400 policymakers on digital inclusion policies for women specifically. So in Pakistan, we worked together with UNESCO on that. And finally, we’re also very pleased, as Ambassador Lange just mentioned, to be one of the implementation partners of the Women in Digital Economy Fund. Thank you.

Leona Verdadero:
Thank you so much. So, so far, it’s been a very rich and enlightening panel discussion, getting perspectives from technology, from government, from civil society and the private sector. We do have a few more questions and we do have time to talk more with our panelists. I have a few questions from myself, but very much happy and eager to open the floor and make it more interactive. So please, if you have a question, raise your hand. We can take a few at a time and then we can take it from there. Yes.

Audience:
I think that, or pardon me, let me circle back. There is a lot of folks talking about access issues, but I rarely hear it framed in terms of managing risk to the people that we get online. How do you think reshaping simply how we. push these messages can positively impact the outcomes on the other end. For instance, are we getting 5,000 more women online? Are we getting 5,000 more women prepared to be online and then getting them online so that way they’re not victimized in the process? And I rarely hear those two things aligned, whether it’s internet access or digitalization. So I’m curious as to your thoughts.

Leona Verdadero:
Thank you. Yes, we’ll note that question. And then if anybody else from the floor has additional questions, we can add that in as well. Okay, for now we can start from there. Yes, if any one of the panelists who would like to address this question first, please go ahead, Ruma.

Rumman Chowdhury:
Thank you for that. I’m nodding even as you ask your question because you’re right, I’ve not heard it framed that way. I will say, you know, and I’ve been at both of the global AI safety summits. So last year there was one at Bletchley, there was just one in Seoul, and this is particularly the one in South Korea was remarkable for its complete lack of presence of global majority. And that was concerning to me. But interestingly, I’m having these conversations here at this summit and most of the answers I’m hearing are along the lines of, well, we’re just in capacity building. We’re just trying to figure out how to even start making AI. We’re not really going to be joining safety conversations because those rooms are, you know, essentially problems of privilege, right? How do we handle with an overwhelming amount of data? You know, but to them, my answer is, well, the reason that that is the conversation is because you are not in the room. I think there is this need to reframe the conversation to talk not just about capacity building and development, but also capacity building for safety and security. And to your point, that also means, you know, online safety for women and girls who would maybe be joining, you know, the internet in particular. I think where the alignment is with how we are talking about AI safety and how, you know, we’re in this room talking about the gender digital divide and also the global majority, is this intersection with mental health and wellbeing. You know, there are countless reports that are demonstrating the adverse impact of artificial intelligence systems, in particular on young people and their impressionable minds, on young women and girls who develop body dysmorphia and negative self-image. And again, just as you’ve mentioned, when we’re saying we’re preparing people, when we are putting people online, what is the online that they get exposed to? There was a paper that I wrote for UNESCO on technology-facilitated gender-based violence and generative AI, and demonstrating how harassment campaigns against prominent women, in particular, you know, journalists and politicians, but really anybody with an online presence, is pretty easy to scale and develop using generative AI. I appreciate your question. I don’t have an answer to it, but I appreciate your question because it frames a problem, not just as a reactive method of addressing something once it happens, but I appreciate it because it’s actually proactive. What is the preparedness that people can have before entering an online space? So, you know, thank you. But also, you know, again, I don’t know if I have an answer, just maybe counter-reflections to your question.

Leona Verdadero:
Yes, please, Anita, go ahead.

Anita Gurumurthy :
Thanks so much. I also think that this is a question of systemic and institutional preparedness. And therefore, while enabling each of us and maybe posterity, our, you know, kids, and, you know, everybody to really be enabled to negotiate this kind of slippery terrain, which also holds a lot of promise, it’s equally important, I think, to think about, you know, thresholds of harm and at what levels will regulation, you know, law and constitutional provisions, jurisprudence, intervene in order to determine what’s not acceptable in our platform cultures. So this is also a question of the production end of algorithms and AI. So what is it that is inadmissible for society in terms of virality and amplification is not a question of preparedness that individual users can answer. This is for the industry to answer. So I think that in terms of regulation, you know, ability of the persuasive skills we have as feminists and as people who care to actually negotiate with industry and be able to say that, you know, this is unacceptable because this is leading to suicides. This is unacceptable. This is leading to mental health crises. So I do think that we should look at this not just only as a question or a solution that is part of preparedness of society, but also the willingness of society to bring in appropriate governance. And I think that’s really important. And of course, institutional capacitation, which may be local policymakers, schools, school teachers as a community. And we find that in my organization as a very important catalyzing force, because teachers are completely unprepared to deal with what comes on the mobile phones of adolescents in school. And they really need and you have to do this on a mass scale. Thank you.

Stephan Lang:
I could just add a few additional thoughts. You know, I think this is an area where we can always do more. From the U.S. government’s perspective and the State Department, we often talk about meaningful access or meaningful inclusion when we talk about digital connectivity. And we’re trying to capture a lot in that phrase. We’re trying to capture the idea that people have the skills they need to be able to use the Internet to be effective, to, you know, make their business more profitable, that they’re safe when they’re online, that they have good cyber hygiene, that they have access to content that’s meaningful to them. whether it’s linguistically or otherwise. And we are trying to do those various things at the same time in different ways. We do have limited capacity, but there are various aspects of trying to make sure that the access that people have is meaningful that we’re trying to address. And rights respecting, too. We have to make sure rights respecting.

Leona Verdadero:
Thank you. Yes, Melle?

Melle Tiel Groenestege:
Yeah, thank you. I like the framing of that question. So we, as GSMA, looked into, let’s say, what we call the user journey. So the process of someone getting online, and particularly for women, what are the risks that women face along that journey? Because they don’t just start at the access point of going online. It’s even before that. So, for example, when you want to do a digital skills training in a community, at what location do you organize those trainings? At what time? Who do you make the trainers? Do women have to share their device to get the training, which may have a privacy risk? If they top up with data, do they need to share their phone number with the sales agent to get the top-up? So we looked at that from a very wide perspective around the potential risks and also what then the industry could do in that regard to address some of those risks. So part of what we do when we train people on a basic digital skill, so this is really for first-time users, we partner with mostly mobile operators, but we’re also open for governments to use our training resources that we have now in 43 different countries, which has a safety component. So this is about very basic knowledge and understanding of what can happen online and how you can change some of your privacy settings or what you should do in case something happens, but much more can be done. And I think, as was reflected here, it requires a multi-stakeholder approach because some of that has to be done by governments in schooling, some of it has to do with policy and regulations, and some of it is on the shoulders of industry and private sector. Thank you.

Leona Verdadero:
We have time for one more question from the room. Yes, please, go ahead.

Audience:
Okay, so I am really grateful for this panel because I think, thankfully from the panelists, we’ve got to hear about global majority voices and from a feminist perspective, just wanted to add another layer to the conversation about youth representation in digital policies. But more particularly, youth is seen as one group, but I wanted to use both of the lenses that the panelists added about women from the global majority, but also youth, because these are the most vulnerable in the groups that we are discussing today. So to frame my question, I wanted to ask you, where do you think young girls from global majority stand in the policymaking process today, and how can they contribute to this conversation? I’m a youth ambassador with Internet Society. I come from Pakistan, so it’s nice to hear this conversation, and I would love to learn how we can contribute and how we can participate in this conversation. Thank you.

Leona Verdadero:
Thank you so much. Yes, who would like to?

Anita Gurumurthy :
Yes, indeed. The UN Secretary General has called for the declaration on future generations, the rights of future generations. I think that’s a very, very important debate, and digitalization, of course, is core to it, as is also climate justice. So that’s one space.

Melle Tiel Groenestege:
I think that’s a good question, and it goes back to, are we addressing as policymakers, or others in that space, the barriers that individuals face? And that is what we try to do as a GSMA, is we try to disaggregate all the data about barriers that people face by location, gender, but also age. And I think that really should be the starting point, is that we have the accurate data and the information. That’s a very top-down approach. Of course, you can also think about more platforms to for youth voices to be raised. But I think taking that user-centric approach in policymaking by focusing on barriers, that should be a starting point for all. Thank you.

Leona Verdadero:
Also just mindful of the time. So we do have three minutes left. And now I would like to call on, actually, to deliver the closing remarks. Cedric Wachols, Chief of Section for Digital Policies and Digital Transformation at UNESCO. Thank you.

Cédric Wachholz:
Good afternoon. I’m delighted to try to sum up a few of the points. Of course, it’s impossible. And I really enjoyed these exchanges. And I think I can speak on the name of our friend and partners, ITU and UNESCO. We are practicing a multi-stakeholder approach with a panel, with an inclusive discussion, even 45 minutes, to the extent possible. And I’m happy about that. Today, we learned and saw how we link the digital inclusion approach and terminology to also digital solidarity in very many concrete examples. But of course, we learned how, secondly, embedding equity and inclusion really needs to happen at the core of policy frameworks and all our digital actions from the outset. And we have learned how diverse inputs are really necessary for fair representation. A lot of discussion about our gender digital divide. For UNESCO, it is one of two global priorities. So we are absolutely delighted about that. Have done a lot of research on that with you, but also with others. And including last March, we published also on large language models, as mentioned. And I really liked and appreciated all. also that, of course, as you said, data of the internet is not the data of the world and the different solutions you offered on that. But of course, when data sets are biased, they produce and reproduce a bias, as we have learned. And we have also heard about the multilayered inequalities and the data misuses and challenges. And I’m not sure if it was all feminist principles, which outlined, because we can discuss that another time. But I mean, I like many of these principles. And I think many of us can endorse them, including empowering inclusion enables local knowledge to be leveraged for collective well-being. And I think there is a strong gender dimension, but it goes even beyond if you take one of those principles. But what I liked also was to hear about a very concrete perspective of user adoption as a fourth point and looking really at the safety, security, affordability dimensions, but also the mental health and the impact on our well-being and dimensions, the concrete work done in these domains. And just to close on a positive note, and also perhaps the questions on how to build global majority to part of a rating, not just building AI systems. That was a good question, too. We’ve heard of all the great work the US is doing in the digital solidarity work, too. And we’re really appreciating that. And we can see it in very many multiple dimensions. The Broadband Commission and co-led enterprise chaired by ITU and UNESCO has done numerous things also for digital inclusion, of course. And we’re holding hands in hands and working together there, of course, with a strong connectivity dimension of ITU and UNESCO bringing in some of the soft dimensions. And to respond also to the youth question, we have a lot of. of youth work in UNESCO ongoing. And it is sometimes a difficulty for us to not just have youth as one of our pillars, but to mainstream it. And we have seen some delegations in UNESCO systematically having youth invited, speak on behalf of their countries and participate very actively in even shaping our programs and budgets for the future. And we have some ways to systematically do that. Generally, for some of the challenges and online challenges provided, we have dual responses on one hand. We have quite recently done the guidelines for regulating online platforms. We’ve done that with the platforms with over 10,000 inputs from all of you and many meetings and so on, but also with media and information literacy to strengthen the capacity, not only of youth, but of the users to detect deep fakes and think before you click and other strategies on dealing with some of our digital challenges. So UNESCO stands for a human rights-based approach for openness, for accessibility and for multi-stakeholder shaped digital environment. And I think the ITU and UNESCO in the Broadband Commission, they look forward also to addressing one key theme. Of course, AI is not our mind, but we saw today that the data governance components is becoming of increasing importance and we’re preparing to launch a working group in this domain jointly soon too, to address some of the challenges today discussed. I thank again all the panelists wholeheartedly for their active contribution and also the participants and online participants and hope you have had a joyful and enriching discussion. Thank you.

AG

Anita Gurumurthy

Speech speed

174 words per minute

Speech length

1374 words

Speech time

474 secs

A

Audience

Speech speed

169 words per minute

Speech length

295 words

Speech time

104 secs

CW

Cédric Wachholz

Speech speed

165 words per minute

Speech length

869 words

Speech time

316 secs

LV

Leona Verdadero

Speech speed

153 words per minute

Speech length

703 words

Speech time

276 secs

MT

Melle Tiel Groenestege

Speech speed

188 words per minute

Speech length

1240 words

Speech time

396 secs

RC

Rumman Chowdhury

Speech speed

195 words per minute

Speech length

1767 words

Speech time

545 secs

SL

Stephan Lang

Speech speed

159 words per minute

Speech length

1213 words

Speech time

458 secs

SA

Sulyna Abdullah

Speech speed

167 words per minute

Speech length

732 words

Speech time

263 secs