WSIS Forum 2026 Inaugural session: High-Level Plenary: The Future We Choose: Leadership for an Inclusive Digital World
The opening plenary of the WSIS Forum 2026 brought together global leaders to discuss digital inclusion, AI governance, and the importance of multilateral cooperation in shaping an equitable digital future . The session was moderated by ITU Secretary General Doreen Bogdan-Martin. .
Annalena Baerbock, President of the 80th session of the UN General Assembly, stressed that multilateralism remains vital but is under pressure, highlighting the WSIS+20 consensus resolution as a significant achievement . She pointed to concrete implementation successes, such as e-health connectivity in Tanzania reaching nearly two million people, while warning that AI governance frameworks are meaningless if basic internet and electricity access remain unavailable to large portions of the global population . President Alar Karis of Estonia outlined five key principles for trusted digital infrastructure: an open and interoperable internet, inclusive multi-stakeholder governance, protection of human rights online, inclusive digital development, and coherence among global digital initiatives . Kazakhstan's Deputy Prime Minister Zhaslan Madiyev detailed his country's rapid digital transformation, including over 90% of public services available online, a constitutional amendment protecting digital rights, and a target to cover one gigawatt of computing capacity within three to five years .
Kailash Satyarthi called for compassion to be embedded in AI development and deployment, arguing that technology is not neutral and must be designed to serve the most vulnerable, particularly children . Bolor-Erdene Battsengel emphasised that AI skilling is an economic imperative, sharing how training teachers as multipliers enabled 50,000 children in remote Mongolian communities to build AI agents . In the rapid-fire round, panellists advocated for a legal right to internet access and big tech liability , universal AI education starting with teachers , treating AI as basic infrastructure like water and electricity , and ensuring engineers engage directly with marginalised communities to build compassion into their work .
Closing remarks from UNESCO, UNCTAD, and UNDP reinforced the need to move from dialogue to implementation, with UNCTAD warning that the geography of the $800 billion AI investment boom is dangerously narrow and that global cooperation is essential to bridge the time gap between infrastructure development and fast-moving AI frontiers . The session concluded with a shared commitment to building inclusive, people-centred digital societies through sustained multi-stakeholder partnership across the WSIS framework and beyond .
Overall Purpose
- The discussion was the opening high-level plenary panel of the WSIS Forum 2026, bringing together heads of state, government ministers, and digital leaders to address how to build an inclusive, people-centred digital future. The central aim was to reflect on the outcomes of the WSIS+20 review process, assess progress on bridging the digital divide, and chart a course for meaningful implementation of digital cooperation commitments - particularly in the age of artificial intelligence.
- --
Major Discussion Points
- The importance of multilateralism and international cooperation in shaping the digital future. Multiple speakers stressed that no single nation can build the digital future alone, and that the WSIS+20 resolution - adopted by consensus - demonstrated that multilateral cooperation remains viable even under pressure. Baerbock warned that this cooperation "is not a given" in the current geopolitical climate and called for cross-regional alliances to connect digital action lines. UNCTAD's representative noted that global cooperation is what can bridge the "time gap" between fast-moving AI development and slower infrastructure-building in developing nations. - Bridging the digital divide, including connectivity, electricity access, and the gap between the Global North and South. Speakers repeatedly highlighted that governance frameworks for AI are meaningless if billions of people lack basic internet access or even electricity. Baerbock specifically cited millions on the African continent who remain unconnected to both the internet and electricity. UNCTAD noted that least developed and lower middle-income countries capture only 10% of strategic sector investments, while $800 billion is being spent on AI infrastructure globally in a single year. Battsengel added that 2.2 billion people still lack access to digital skills and the internet - The principles of trusted, inclusive digital public infrastructure and governance. President Karis of Estonia outlined five key principles for building trusted digital infrastructure: keeping the internet open, secure, and interoperable; ensuring inclusive multi-stakeholder governance; protecting human rights online; advancing inclusive digital development; and ensuring that global digital initiatives reinforce rather than duplicate each other. These principles were summarised by the moderator, as a framework for the broader discussion. - AI skilling, education, and the need to build human capacity at scale. Several panellists emphasised that connectivity alone is insufficient without the skills to use it meaningfully. Battsengel described how training teachers as "multipliers" enabled 50,000 children in rural Mongolia to build AI agents. Kazakhstan's Deputy Prime Minister reported training one million people in AI skills in a single year and announced the launch of an AI University. Estonia's President described a national programme ensuring 100% of teachers and students are engaged with AI education. Battsengel argued that building AI skills is now equivalent to "building an economic engine." - Child safety, compassionate AI, and the moral responsibility of technology builders. Nobel Peace Laureate Kailash Satyarthi made a powerful case that technology is not neutral and that AI systems carry the values of those who build them. He called for "compassionate AI" - embedding justice, equality, peace, and sustainability into the development and deployment of AI algorithms. His concrete proposal was for the tech industry to send engineers to spend one month with vulnerable children in mines, war zones, and remote villages, arguing this experience would fundamentally change how they write code. The moderator highlighted child safety as a key theme from the preceding Global Dialogue on AI Governance. ---
Overall Tone
- The overall tone of the discussion was earnest, collaborative, and cautiously optimistic, with moments of urgency and moral gravity. The opening remarks established a spirit of collective purpose, with the moderator framing the session around leadership, trust, and inclusion. As the panel progressed, the tone became more candid and at times sobering, particularly when Baerbock acknowledged that multilateralism and international cooperation are "under heavy attack" and that even basic rights such as housing and education are endangered by the digital divide. Satyarthi introduced the most emotionally charged moments, shifting the tone towards moral urgency, drawing on his decades of advocacy for children's rights to challenge the room to act with compassion.
- The closing contributions from UNESCO, UNCTAD, and UNDP returned the tone to one of institutional commitment and forward-looking resolve, with speakers acknowledging the scale of the challenge while reaffirming their organisations' dedication to implementation.
Opening Plenary of the WSIS Forum 2026: Leadership, Inclusion, and the Digital Future
#
Introduction and Context
The opening high-level plenary of the WSIS Forum 2026 brought together an exceptional group of global leaders to address one of the defining challenges of the contemporary era: how to build a digital future that is inclusive, equitable, and genuinely beneficial to all people, everywhere. The session was moderated by ITU Secretary General Doreen Bogdan-Martin, who framed the central question as one of leadership - "leadership that inspires action, that builds trust, and delivers results for a more inclusive digital world" . The panel featured Her Excellency Annalena Baerbock, President of the 80th Session of the United Nations General Assembly; His Excellency Alar Karis, President of the Republic of Estonia; His Excellency Zhaslan Madiyev, Deputy Prime Minister and Minister of Artificial Intelligence and Digital Development of Kazakhstan; Mr Kailash Satyarthi, Nobel Peace Laureate and founder of the Satyarthi Movement for Global Compassion; and Ms Bolor-Erdene Battsengel, Chief Executive Officer of AI Academy Asia . Notably, Egypt assumed the chair of the WSIS Forum 2026, a detail acknowledged in the opening remarks. The discussion unfolded against the backdrop of the recently concluded WSIS Plus 20 review process and the Global Dialogue on AI Governance, lending it both historical weight and immediate policy relevance .
#
The State of Multilateralism and the WSIS Plus 20 Achievement
The session opened with a frank assessment of the state of multilateralism from Annalena Baerbock, who acknowledged that international cooperation, once taken for granted, is now "under heavy attack" . She argued that this fragility makes the WSIS Plus 20 consensus resolution all the more significant - describing it as one of the rare moments where the General Assembly reached consensus, a formulation that in the transcript appears in garbled form but clearly conveys a demonstration that the international community can still act collectively . Baerbock was careful to contextualise this achievement within a broader geopolitical reality, noting that even the understanding of what institutions such as the ITU do on a daily basis "is not a given" in the current climate . She also invoked the three pillars of the United Nations - peace and security, human rights, and development - as heavily interconnected, arguing that bridging the digital divide across gender lines and between the Global North and Global South requires addressing all three simultaneously. Her call was for individual leaders from different continents to form cross-regional alliances to connect the WSIS action lines and ensure that those still excluded from digital progress are not left further behind .
This affirmation of multilateralism was echoed throughout the session. President Karis of Estonia stated that the digital future "is not something that simply happens to us - it is a choice we make, the principles we defend, the cooperation we build" . Robert Opp, UNDP's Chief Digital Officer, described the WSIS Plus 20 review as having "reaffirmed the WSIS vision that has been remarkably prescient for the last 20 years of an inclusive, people-centred and development-oriented digital future" . UNESCO's Mariya Gabriel, Assistant Director General for Communications and Information, affirmed that "the WSIS architecture holds, the action lines hold, the multi-stakeholder processes hold," and that the Global Digital Compact has extended the framework's function as an implementation engine . Across all speakers, there was a shared conviction that the priority must now shift decisively from reflection and dialogue to concrete implementation and delivery of results .
#
Bridging the Digital Divide: Foundational Infrastructure and the Limits of AI Governance
A recurring and urgent theme throughout the session was the persistent and potentially widening gap between those with access to digital infrastructure and those without. Baerbock made the point with particular force, arguing that advanced AI governance frameworks are entirely meaningless if large portions of the global population lack basic internet access or even electricity . She cited the specific example of millions in sub-Saharan Africa who remain unconnected not only to the internet but to the electricity grid itself, making AI-driven health solutions irrelevant in the absence of foundational infrastructure . She also pointed to concrete implementation successes under the WSIS action lines - such as the e-health connectivity project in Tanzania under Action 9-7, which connected almost two million people to between twenty and sixty health facilities - as evidence of what is possible when commitments are translated into action .
Pedro Manuel Moreno, Deputy Secretary-General and Acting Secretary-General of UNCTAD, provided the most analytically rigorous framing of this structural inequality. He reported that the world will spend $800 billion on AI infrastructure in the current year alone, yet the geography of this investment boom is "dangerously narrow," with least developed and lower middle-income countries capturing only 10% of strategic sector investments . He introduced the concept of a "time gap" - a mismatch between the rapid pace of AI development, which runs on quarterly cycles, and the long construction timelines for infrastructure in developing nations: a data centre takes three years to build, a power plant five, a grid ten, and an education system a generation . His sobering conclusion was that even under the most optimistic scenario, with the right investments and policies, "building alone gets most countries to the table after the rules are written" . Global cooperation, he argued, is what can bridge this time gap, and multilateral forums such as WSIS are the mechanism through which developing nations can shape the rules of the digital economy before they are finalised .
Battsengel reinforced this picture from the perspective of lived experience, referencing ITU reports indicating that 2.2 billion people around the world have not been able to access digital skills or the internet . She framed this not merely as a development challenge but as a violation of human rights - "it is blocking them, it's violating their human rights to learn, to get equal education" . She also made the pointed observation that many technology and AI companies are not participating in discussions on AI education, AI human rights, and AI inclusion, and called for them to be brought to the table and actively involved in these conversations. Gabriel of UNESCO added that billions cannot access the internet, and that millions of teachers and learners are navigating digitalisation with tools "not built for their languages or their realities" , underscoring that the digital divide has linguistic and cultural dimensions as well as infrastructural ones. Gabriel also noted that young people - who account for 60% of Africa's population - cannot be left out of these conversations if the international community is to be true to its mission of shaping a better future.
#
Five Principles for Trusted Digital Public Infrastructure: Estonia's Framework
President Karis of Estonia offered the most structured governance contribution of the panel, outlining five key principles that he argued should guide the building of trusted, inclusive digital public infrastructure. Drawing on Estonia's more than two decades of experience in digital transformation, he stated that these principles had been validated through practice rather than theory . (It should be noted that the transcript contains an apparent transcription error at this point, rendering the country name as "Germany" rather than "Estonia"; the context makes clear that Karis was speaking about Estonia's own experience.)
The first principle was that the internet must remain open, secure, and interoperable, with Karis warning that fragmentation and excessive centralisation would weaken both innovation and the democratic potential of technology . The second was that digital governance must be inclusive of all stakeholders - governments, the private sector, civil society, the technical community, and academia - since governments "cannot govern the digital world alone" . The third was that human rights must be protected online just as they are offline, with AI remaining "human-centred, rights-based and accountable" . The fourth was that digital development must be inclusive, encompassing not only access but also digital literacy, data literacy, and the ability to understand how automated systems work . Karis added a cultural dimension here, arguing that AI systems must work not only in the largest global languages but also in local languages and cultural contexts . The fifth and final principle was that global digital initiatives - from the WSIS to the Global Digital Compact - must reinforce each other rather than create duplication . These principles were explicitly summarised and endorsed by the moderator as a framework for the broader discussion .
#
Kazakhstan's Digital Transformation: A Model of Ambition and Scale
Deputy Prime Minister Zhaslan Madiyev of Kazakhstan provided a detailed account of his country's rapid and comprehensive digital transformation, describing a decade of substantial investment that has positioned Kazakhstan among the leading nations in e-government development . He outlined a people-centric approach in which over 90% of more than 1,300 public services are available online, with services built around life situations rather than bureaucratic processes - enabling citizens to buy or sell a car in two to three minutes, apply for a mortgage and receive approval within a day, and automatically register a newborn child's identity . More than 90% of transactions in Kazakhstan are now cashless, and the National Bank is piloting a tokenised, programmable central bank digital currency .
Madiyev described Kazakhstan's broader digital ecosystem, including the Astana Hub - one of the leading IT hubs in the region - which hosts more than 2,000 resident IT companies and extends its reach to Silicon Valley, Dubai, Shanghai, and beyond . He noted that every ministry in the government has a dedicated digital deputy minister, illustrating the depth of Kazakhstan's institutional commitment to digital transformation. He outlined an ambitious legislative agenda undertaken in the first half of the current year, including the approval of a law on AI, a digital code, and - most strikingly - a constitutional amendment embedding digital rights and personal data protection directly into Kazakhstan's constitution . He also described the launch of an AI Council under the President, comprising prominent figures from around the world, as well as the establishment of ALAM AI, an International AI Centre designed to build a fully integrated innovation ecosystem encompassing large technology companies, startups, school children, and students. Looking ahead, he described a target to build one gigawatt of computing capacity within three to five years , and reported that Kazakhstan covered one million people with AI skills in the previous year, with an AI University set to launch in September of the current year . In the rapid-fire round, Madiyev proposed that AI should be treated as base infrastructure for people - equivalent to water supply, electricity, and the internet - arguing that this framing is essential to prevent the widening of inequality . He added a second priority: changing the mindset of youth to embrace entrepreneurship, with a vision of merging innovation ecosystems with universities so that students can receive academic credit for incubation and acceleration programmes.
#
Child Safety, Compassionate AI, and the Moral Responsibility of Technology Builders
The most emotionally charged and philosophically distinctive contribution came from Nobel Peace Laureate Kailash Satyarthi, who reframed the entire discussion around moral responsibility and human compassion. Drawing on his decades of advocacy for children's rights - including the Global March Against Child Labour that crossed 103 countries and engaged 7.2 million people, ultimately leading to the unanimous adoption of the ILO Convention on the Worst Forms of Child Labour - he argued that he had returned to Geneva because "the rules for the future of human intelligence are being written here" .
Satyarthi's central argument was that technology is not neutral: "they carry the minds and purpose and intentions and values of those people who create them" . He illustrated this with a striking analogy: just as people walk across a bridge without questioning its safety, trusting that those who built it accepted a moral responsibility to make it safe, so too must the builders of AI systems accept a moral responsibility for the safety and wellbeing of those who will depend upon them. He challenged the room with a fundamental question - not what AI systems can do, but "who these systems are going to serve and protect" and "whether they are going to protect and serve the last child of society, the most vulnerable people, or going to exploit them more" . He called for compassion to be embedded in AI development and deployment not as "a soft emotion" but as "a disruptive, dynamic and transformative force" encompassing justice, equality, peace, and sustainability . His concrete proposal in the rapid-fire round was striking in its specificity: he called on industry to give their engineers one month to meet children in mines and factories, in war zones, and in remote villages, arguing that this experiential encounter would fundamentally change how they design AI systems . "When they come back to their labs, they would be different individuals," he said. "This one month can shape the future of humankind" .
This proposal generated the most immediate and visible response from a fellow panellist. Battsengel explicitly stated she would "be happy to walk with you" and would give her own engineers for one month as well , demonstrating that the proposal resonated as practically viable rather than merely inspirational. The convergence of a moral activist and a technology practitioner on this concrete, actionable proposal for humanising AI development was one of the most unexpected and significant moments of consensus in the session.
#
AI Skilling as Economic Infrastructure: The Multiplier Effect of Teacher Training
Battsengel's contribution was distinctive for its combination of concrete empirical evidence and a reframing of AI skilling as economic infrastructure rather than charitable access provision. She opened with the story of a girl from a nomadic family who had never stepped into a city, never used a computer, and never seen a smart device, who - after being reached by the Girls' Code NGO - went on to graduate from an Ivy League university on a full scholarship and is now working in New York . This narrative served as a powerful anchor for her central argument: that when access and equal educational opportunity are provided, transformation is achievable.
Battsengel argued that the nature of education itself has changed in the age of AI: "we don't need people who have knowledge - we need people who have skills," and that "building AI skills no longer means training - it means building an economic engine, it means creating competitiveness, it means creating opportunities and filling gaps" . She reported that AI Academy Asia trained 15,000 people in Mongolia and Central Asia in the previous year and is planning to train 100,000 people across Africa and Asia in the current year . The most operationally significant insight she shared was the importance of training teachers as multipliers: "teachers are the multipliers in the education system" . She described how training 1,000 teachers from 330 villages in Mongolia led to those teachers training 50,000 children - children "coming from riding horses, getting raised in the wilderness" - who were subsequently able to build their own AI agents . This evidence-based argument for a teacher-first strategy was independently echoed by President Karis, who described how Estonia had similarly started its AI education programme by training teachers, with 100% of teachers and students in upper secondary schools now involved .
#
The Rapid-Fire Round: Divergent Priorities, Shared Urgency
The second round of questions, moderated by Sulyna Abdullah, invited each panellist to identify the single change they would make to accelerate universal meaningful connectivity and digital inclusion . The responses revealed meaningful divergences in priorities and conceptual framings, even within a broadly shared normative framework.
Baerbock called for a formal right to internet access, acknowledging this was controversial but arguing that without it, other fundamental human rights - including housing, education, and economic participation - are endangered . She cited the experience of COVID-19, when students without home internet access were effectively denied their right to education . She added a second priority: establishing liability for big tech companies, drawing an explicit analogy with nuclear technology and arguing that "if we do not do the same with the power of AI, all the human rights and not only education, but the human rights of human dignity are in danger" . Karis returned consistently to education as the cornerstone, describing Estonia's progression from the "tiger leap" of providing internet access to all schools to the current "AI leap" integrating AI literacy for all teachers and students . Madiyev proposed treating AI as base infrastructure equivalent to water and electricity, and added a second priority: changing the mindset of youth to embrace entrepreneurship, with a vision of merging innovation ecosystems with universities so that students can receive academic credit for incubation and acceleration programmes . Satyarthi's answer was entirely different in kind - rather than a regulatory or institutional change, he proposed a change in human experience and empathy, calling for engineers to spend a month with marginalised children . Battsengel focused on ensuring that connectivity is accompanied by meaningful digital learning platforms, emphasising that "digital inclusion doesn't mean using social media, Instagram, TikTok - it means using that to get more opportunities to learn, to embrace the technology and fill the gap" .
#
Closing Contributions: From Dialogue to Delivery
The closing segment brought together the co-organising institutions - ITU, UNESCO, UNCTAD, and UNDP - to offer institutional perspectives and commitments. Gabriel of UNESCO affirmed that "WSIS was always about people, not technology" and that the vision of information and knowledge societies built on human rights, cultural diversity, and the free flow of knowledge "has never been more necessary" . She noted that AI sits on a foundation that too many people - particularly in the Global South and rural areas - cannot yet rely upon, and committed UNESCO to investing in low-resource AI tools, edge computing that works without data centres, and digital competency frameworks designed for teachers everywhere .
Moreno of UNCTAD, in addition to his analysis of the investment gap and time divide, reported that UNCTAD had chaired the UN Group on the Information Society, which delivered a joint implementation roadmap to align WSIS and the Global Digital Compact, and had launched a task force on financing for digital development aimed at addressing the investment gap facing developing countries . He concluded with a formulation that the moderator signalled she would "plagiarise": "20 years ago the summit made connectivity a shared project - our task now is to make the digital economy a shared" . The sentence appears incomplete in the transcript rather than being a deliberate rhetorical device, but the aspiration it conveys is unmistakable.
Opp of UNDP stressed that the forum must move beyond reflection to implementation, with the next decade focused on "delivering results together" . He cited UNDP's work across 26 countries on AI transformation, noting that the key question is not whether countries are AI-ready but "whether they're building the capability to steer AI toward public value" . He described concrete examples, including support for national AI strategies in Zimbabwe and Ghana paired with training for policymakers , and the use of public procurement as a practical governance lever for managing AI adoption in Trinidad and Tobago, often ahead of formal legislation . He concluded by affirming that "multi-stakeholder partnership is not simply support for implementation - it's the implementation model itself," requiring government, the UN system, the private sector, academia, civil society, the technical community, and young innovators, "each bringing something that no single institution can possess" .
#
Areas of Consensus and Unresolved Tensions
The session was characterised by a high degree of surface-level consensus around shared goals - inclusive digital development, bridging the digital divide, people-centred AI, and the continued relevance of the WSIS framework - but revealed meaningful divergences in priorities and mechanisms when speakers were pressed for specific recommendations.
The most substantive areas of agreement included: the enduring validity of the WSIS framework and the need to shift from dialogue to implementation ; the centrality of education and teacher training as the most effective lever for digital and AI inclusion ; the insufficiency of connectivity alone without skills and meaningful access ; the imperative that AI be human-centred, rights-based, and serve the most vulnerable ; and the necessity of multi-stakeholder partnership as the implementation model itself .
However, genuine tensions remained unresolved. Baerbock's call for legal liability for big tech companies found no direct endorsement from other panellists, who framed the private sector primarily as a partner to be engaged rather than an actor to be regulated - an observation that reflects the analytical divergence in the session rather than an explicit statement made by any speaker. The question of whether AI is primarily an infrastructure and economic challenge - as Madiyev framed it - or a moral and ethical one requiring compassion-driven design - as Satyarthi argued - reflected genuinely different conceptual framings with significant implications for governance. The protection of undersea cable infrastructure from deliberate attack was raised by Baerbock as an emerging security concern but was not addressed by any other panellist, leaving the peace and security dimension of digital governance largely unexplored. And the structural challenge identified by Moreno - that developing countries risk arriving at the AI governance table after the rules have already been written - was acknowledged as urgent but left without a concrete resolution.
#
Conclusion
The opening plenary of the WSIS Forum 2026 demonstrated both the maturity of the international community's shared understanding of the challenges of inclusive digital development and the scale of the work that remains to be done. The session moved from celebration of the WSIS Plus 20 consensus to a more demanding and urgent conversation about implementation, accountability, and the structural barriers - financial, temporal, linguistic, and moral - that stand between current commitments and genuine digital inclusion for all. Abdullah closed the session by invoking the transition "from dialogue to delivery and from the information society to the intelligence society" , capturing the collective aspiration of a forum that has spent two decades building the architecture for a people-centred digital world and must now demonstrate that it can deliver on that promise.
Multilateralism is alive but under pressure - defending international cooperation is critical in fragmented times
Arg. 1Baerbock argues that while multilateralism remains alive, it can no longer be taken for granted as the international order faces heavy attack. In these fragmented and challenging times, the commitment to building the future together through cooperation is more crucial than ever.
She noted that a few years ago affirming multilateralism would have been unnecessary, but today the international order and cooperation are not only under pressure but under heavy attack . She emphasised that the understanding of what ITU does on a daily basis in terms of connectivity is no longer a given .
The WSIS Plus 20 resolution adopted by consensus represents a rare and powerful moment of global agreement
Arg. 2Baerbock highlights the WSIS Plus 20 resolution as a significant achievement because it was adopted by consensus, which is increasingly rare in the current geopolitical climate. She argues this consensus should be leveraged as a source of power for advancing digital cooperation and governance.
She specifically referenced the joint resolution and WSIS Plus 20 adopted in consensus the previous winter, describing it as one of the few moments where the General Assembly reached consensus . She called this the power that should be used in the fields of technical cooperation, digital cooperation, and governance .
Billions remain unconnected, and without basic electricity and internet access, advanced technologies like AI are meaningless for the most vulnerable
Arg. 3Baerbock argues that advanced digital governance structures such as AI frameworks are futile if large portions of the population lack basic internet access or even electricity. She stresses that connectivity must be addressed as a foundational prerequisite before higher-level digital solutions can have any meaningful impact.
She pointed out that if a fifth or even a quarter of the population does not have internet access, an AI model for e-health is useless if basic services are not in place . She specifically cited the African continent and sub-Saharan Africa, where millions are not connected to the internet or even to electricity, making AI health centres irrelevant .
on: The digital divide — particularly between the global north and south, and between connected and unconnected populations — remains a critical and urgent challenge
A right to internet access should be formally guaranteed, as its absence endangers other fundamental human rights including housing and education
Arg. 4Baerbock argues that internet access should be recognised as a formal right because the absence of digital access increasingly undermines other established human rights. She contends that as essential services move online, those without internet access are effectively excluded from exercising rights to housing, banking, and education.
She noted that in many countries it is now impossible to rent an apartment without a bank account, and digital banking is increasingly required to access banking services . She also referenced the COVID-19 pandemic, during which students without home internet access were denied their right to education, including in her own city near Berlin .
on: Whether legal and rights-based frameworks or practical capacity-building should be the primary mechanism for advancing digital inclusion
Liability for big tech companies must be established to protect human dignity, just as liability was established for other powerful technologies such as nuclear energy
Arg. 5Baerbock argues that the immense power held by big tech companies in the AI era must be matched with legal accountability, drawing a parallel with how liability frameworks were established for nuclear technology. Without such accountability, fundamental human rights including human dignity are at risk.
She drew a direct analogy with nuclear technology, arguing that when revolutionary technological changes occurred in the past, societies ensured liability for those who held that power . She concluded that without equivalent liability for AI, human rights and human dignity are endangered .
on: AI must be human-centred, rights-based, and designed to serve and protect the most vulnerable members of society
on: Whether accountability for big tech companies through liability frameworks is a necessary component of digital governance
Individual leaders from different continents must form cross-regional alliances to connect WSIS action lines and address those still excluded from digital progress
Arg. 6Baerbock calls for individual leadership and cross-regional alliances as essential mechanisms for connecting the various WSIS action lines and ensuring that digital progress reaches those still excluded. She argues that without such coordinated leadership, the gains from international agreements will not translate into meaningful outcomes for the most vulnerable.
She cited concrete examples of successful WSIS action line implementation, including Action 9-2 on infrastructure in Peru and Action 9-7 on e-health in Tanzania, where almost 2 million people were connected to 20 to 60 health facilities . She called for leaders from different continents to form cross-regional alliances to connect these action lines .
The three pillars of the United Nations — development, human rights, and peace and security — are deeply interconnected and must be addressed together in digital governance
Arg. 7Baerbock argues that digital governance cannot be siloed and must be understood in relation to all three pillars of the UN system: development, human rights, and peace and security. Failure to bridge these pillars will undermine the overall success of digital transformation efforts.
She highlighted the vulnerability of undersea internet cables as an example of the peace and security dimension of digital infrastructure, noting that without regulation, container ships could attack these cables with no protection in place . She concluded that bridging the digital divide across gender, the global north-south divide, and the three UN pillars is essential to success .
The WSIS framework has proven remarkably prescient over 20 years and must now shift from reflection to implementation
Arg. 1Karis acknowledges that the World Summit on Information Society raised the right questions about ensuring digital transformation serves people and supports development for all, and that these questions remain as relevant today as they were two decades ago. He argues that the focus must now move from articulating principles to delivering concrete results.
He noted that more than 20 years ago the World Summit on Information Society changed things profoundly, yet the central question of how digital transformation serves people and supports development for all remains the same . He stated that for Estonia, this is not an abstract question, as the country has built much of its modern state on a digital foundation .
on: The WSIS framework has proven its value and must now shift decisively from dialogue and reflection to concrete implementation
Global initiatives such as WSIS and the Global Digital Compact must reinforce each other rather than create duplication
Arg. 2Karis argues that the proliferation of global digital governance processes risks creating confusion and inefficiency if they do not actively reinforce one another. He calls for coherence between the WSIS process, the Global Digital Compact, and other international initiatives.
He explicitly stated as his fifth principle that the WSIS process and other initiatives should reinforce each other and not create duplication, and highlighted the important role of international organisations such as ITU in setting standards, building capacity, and supporting interoperability .
The internet must remain open, secure, and interoperable, as fragmentation weakens both innovation and democratic potential
Arg. 3Karis contends that the internet's fundamental strength lies in not being controlled by any single state, company, or institution, and that fragmentation or excessive centralisation would undermine both innovation and democracy. He presents this as the first of five key principles for trusted digital public infrastructure.
He stated that fragmentation, excessive centralisation, and closed technology spheres would weaken innovation and the democratic potential of technology . He noted that the internet supports education, entrepreneurship, public debate, scientific cooperation, and freedom of expression .
Digital governance must be inclusive of all stakeholders — governments, private sector, civil society, academia, and the technical community
Arg. 4Karis argues that while governments have a crucial role in setting rules and protecting rights, they cannot govern the digital world alone and must work alongside the private sector, civil society, the technical community, and academia. He presents inclusive multi-stakeholder governance as the second of his five key principles.
He stated that governments have a crucial role in setting rules and protecting rights but cannot govern the digital world alone, and that the private sector, civil society, the technical community, and academia all have a central role to play . He also referenced the Digital Inclusion and Governance Forum as a body that brings stakeholders together as equals .
on: Multi-stakeholder partnership — encompassing governments, private sector, civil society, academia, and the technical community — is essential for implementing the digital future
on: Whether accountability for big tech companies through liability frameworks is a necessary component of digital governance
Human rights must be protected online just as offline, and AI must remain human-centred, rights-based, and accountable
Arg. 5Karis presents the protection of human rights online as his third key principle, arguing that this is especially critical in the age of artificial intelligence. He stresses that AI must be developed and deployed in a manner that is human-centred, rights-based, and accountable to ensure it strengthens rather than undermines democratic societies.
He referenced Estonia's participation in the Freedom Online Coalition in 2025, where Estonia stood for an internet that supports democracy, transparency, and trust . He stated that AI must remain human-centred, rights-based, and accountable, and that digital transformation can strengthen societies only when it allows for civic space and democratic accountability .
on: AI must be human-centred, rights-based, and designed to serve and protect the most vulnerable members of society
Education is the single most important factor for digital transformation; Estonia prioritised internet access in all schools and is now integrating AI literacy for all teachers and students
Arg. 6Karis argues that education has been and remains the cornerstone of Estonia's digital transformation journey, from the early provision of internet access in all schools to the current integration of AI literacy. He presents this as the key lesson from Estonia's experience that other countries should adopt.
He described Estonia's 'tiger leap' initiative from 20 years ago, which provided internet access to all schools across the country, not just in big cities, creating the foundation for the current 'AI leap' . He noted that in the second year of Estonia's AI education programme, 100% of teachers and 100% of students are involved .
on: Education — particularly teacher training — is the most effective and scalable lever for digital and AI inclusion
on: Whether legal and rights-based frameworks or practical capacity-building should be the primary mechanism for advancing digital inclusion
Kazakhstan has built a people-centric e-government system with over 90% of public services available online, demonstrating the value of digital public infrastructure
Arg. 1Madiyev presents Kazakhstan's e-government achievements as evidence that sustained investment in digital public infrastructure can deliver tangible, people-centred results. He argues that services built around life situations and accessible online represent the kind of inclusive digital transformation that other nations should aspire to.
He stated that Kazakhstan has entered the list of top leading nations on e-government development, with more than 90% of over 1,300 public services available online . He provided concrete examples such as buying or selling a car in two to three minutes with home delivery of the car plate, applying for a mortgage and receiving approval within a day, and automatic government registration upon the birth of a child .
Kazakhstan has embedded digital rights and personal data protection directly into its constitution, reflecting a comprehensive legislative approach to digitalisation
Arg. 2Madiyev highlights Kazakhstan's constitutional amendment protecting digital rights and personal data as a landmark step in creating a comprehensive legal framework for digitalisation. He argues that this legislative approach, alongside a new AI law and digital code, reflects the depth of Kazakhstan's commitment to building a trusted digital society.
He stated that Kazakhstan amended its constitution to preserve digital rights and personal data protection within it . He also noted the approval of a law on AI, a digital code, and upgrades to the law on telecommunications and data centres, including cloud-first policies .
on: Whether legal and rights-based frameworks or practical capacity-building should be the primary mechanism for advancing digital inclusion
AI should be considered base infrastructure for people, equivalent to water supply and electricity, to prevent widening inequality
Arg. 3Madiyev argues that AI should be elevated to the status of essential infrastructure, comparable to water and electricity, to ensure it is universally accessible and does not exacerbate existing inequalities. He contends that treating AI as base infrastructure is the key to ensuring it delivers productivity gains and efficiencies for all.
He stated that AI should be the base infrastructure for people going forward, like water supply, electricity, and internet, because it will allow societies to avoid gaps by providing productivity gains and efficiencies .
on: Whether AI should be framed primarily as infrastructure/economic tool or as a moral and ethical challenge requiring compassion-driven design
Kazakhstan covered one million people with AI skills in one year and is launching an AI University, reflecting the scale of investment needed in human capital
Arg. 4Madiyev presents Kazakhstan's large-scale AI skills programme as evidence of the ambition and investment required to build the human capital necessary for a digital nation. He argues that human capital development must cover all population groups, from school children to government employees.
He stated that last year Kazakhstan covered one million people with AI skills , and that an AI University is to be launched in September of the current year . He also described the ALAM AI International AI Centre as a platform covering broad population groups, big tech companies, startups, school children, and students .
on: Education — particularly teacher training — is the most effective and scalable lever for digital and AI inclusion
Fostering an entrepreneurial mindset among youth — encouraging them to build startups rather than seek public sector employment — is essential for building a technological nation
Arg. 5Madiyev argues that a cultural shift in how young people view their career aspirations is as important as technical skills development for building a truly digital and innovative nation. He calls for merging the innovation startup ecosystem with universities so that students are incentivised to become founders rather than public sector employees.
He stated that Kazakhstan is planning to merge its innovation startup ecosystem with universities so that students can receive academic credit for acceleration and incubation programmes . He expressed the aspiration that everyone should dream of building a unicorn company in their lifetime .
Technology is not neutral; it carries the values and intentions of its creators, and leaders must ensure AI serves and protects the most vulnerable, including children
Arg. 1Satyarthi argues that AI systems are not neutral tools but reflect the minds, purposes, and values of those who create them, making it imperative that leaders ensure these systems serve and protect the most vulnerable members of society. He frames this as a fundamental moral responsibility of those who build and deploy AI.
He used the analogy of a bridge, arguing that those who build a bridge have a moral responsibility to ensure it is safe for people to walk on, and that the same principle applies to those who build AI systems . He stated that technology carries the minds, purposes, intentions, and values of those who create it , and asked whether AI systems will protect the last child of society or exploit them further .
on: AI must be human-centred, rights-based, and designed to serve and protect the most vulnerable members of society
on: Whether AI should be framed primarily as infrastructure/economic tool or as a moral and ethical challenge requiring compassion-driven design
Compassion must be embedded as a transformative force — encompassing justice, equality, peace, and sustainability — in the development and deployment of AI algorithms
Arg. 2Satyarthi calls for compassion to be understood not as a soft emotion but as a disruptive, dynamic, and transformative force that must be embedded in the technical development and deployment of AI. He argues that compassion, encompassing justice, equality, peace, and sustainability, must be built into AI algorithms and digital systems.
He explicitly called for compassion as a disruptive, dynamic, and transformative force with its four elements of justice, equality, peace, and sustainability to be embedded in the development and deployment of AI algorithms and digital systems . He framed this as 'compassionate AI' that reaches out to the last person in society .
on: Whether AI should be framed primarily as infrastructure/economic tool or as a moral and ethical challenge requiring compassion-driven design
Engineers and AI developers should spend time with marginalised children to understand the human consequences of the code they write, fostering empathy-driven design
Arg. 3Satyarthi proposes that industry send their engineers to spend one month with marginalised children in mines, factories, war zones, and remote villages, arguing that this direct human experience would fundamentally change how they approach their work. He contends that when developers connect with and feel the realities of those affected by their code, they will return to their labs as different individuals who design with greater compassion.
He drew on his own experience of the Global March Against Child Labour, which crossed 103 countries and involved 7.2 million people, leading to the unanimous adoption of the ILO Convention on the Worst Forms of Child Labour . He invited engineers to walk with him in villages where children live with uncertainty and hopelessness, promising that this one month could shape the future of humankind .
on: The single most important intervention to accelerate universal meaningful connectivity and digital inclusion
Digital inclusion must go beyond connectivity to ensure meaningful access, including digital skills and learning platforms for last-mile communities
Arg. 1Battsengel argues that simply providing connectivity is insufficient for true digital inclusion; what matters is ensuring that people can use that connectivity to learn, create opportunities, and develop skills. She stresses that digital inclusion means access to digital platforms for learning and AI tools, not merely access to social media.
She shared the story of a girl from a nomadic family living 2,000 kilometres from the city who, through the Girls' Code NGO, gained access to education and eventually graduated from an Ivy League school with a full scholarship . She stated that digital inclusion does not mean using social media like Instagram or TikTok, but using connectivity to get more opportunities to learn and embrace technology .
on: The digital divide — particularly between the global north and south, and between connected and unconnected populations — remains a critical and urgent challenge
on: The single most important intervention to accelerate universal meaningful connectivity and digital inclusion
Building AI skills is no longer just about training; it is about building economic engines, creating competitiveness, and filling opportunity gaps
Arg. 2Battsengel argues that the nature of AI skills development has fundamentally changed the purpose of education, shifting from the accumulation of knowledge to the development of practical skills. She contends that AI skilling is now an economic imperative that drives competitiveness and creates opportunities.
She argued that traditional education systems built people with wide or specific knowledge, but AI has changed this because what is now needed is people with skills rather than generic knowledge . She stated that building AI skills means building an economic engine, creating competitiveness, and filling gaps .
on: Whether legal and rights-based frameworks or practical capacity-building should be the primary mechanism for advancing digital inclusion
Training teachers as multipliers is more effective than focusing solely on students; 1,000 trained teachers in Mongolia reached 50,000 children in rural communities
Arg. 3Battsengel argues from direct experience that training teachers is the most effective strategy for scaling AI education, as teachers act as multipliers who can reach far more children than direct student training programmes. She presents this as a key lesson for government leaders designing AI education strategies.
She described how AI Academy Asia trained 1,000 teachers from public schools in 330 villages in Mongolia, and those teachers subsequently trained 50,000 children . She noted that these 50,000 children, many from rural backgrounds, were able to build their own AI agents and use AI to solve problems in their communities .
on: Education — particularly teacher training — is the most effective and scalable lever for digital and AI inclusion
Denying children and last-mile communities access to digital education and AI tools is a violation of their human rights to learn and develop equally
Arg. 4Battsengel frames the lack of access to digital skills and internet connectivity not merely as a development gap but as a violation of fundamental human rights, specifically the right to equal education. She argues that the 2.2 billion people without access to digital skills and internet are being actively blocked from exercising their rights.
She referenced ITU reports indicating that 2.2 billion people around the world have not been able to access digital skills or the internet . She stated explicitly that this is not only about giving them connectivity but that blocking access violates their human rights to learn and receive equal education .
on: AI must be human-centred, rights-based, and designed to serve and protect the most vulnerable members of society
Tech companies and AI firms must be brought to the table on AI education, inclusion, and human rights discussions, as many are currently absent from these conversations
Arg. 5Battsengel argues that the private sector, particularly tech and AI companies, must be actively engaged in discussions about AI education, inclusion, and human rights, as their current absence from these conversations is a significant gap. She calls for industry to collaborate more effectively and be involved in these critical topics.
She stated that many tech companies and AI companies are not taking part in AI education, AI human rights, and AI inclusion discussions, and called for them to be brought to the table and involved in these topics .
on: Multi-stakeholder partnership — encompassing governments, private sector, civil society, academia, and the technical community — is essential for implementing the digital future
on: Whether accountability for big tech companies through liability frameworks is a necessary component of digital governance
The WSIS architecture, action lines, and multi-stakeholder processes remain valid and have been extended by the Global Digital Compact
Arg. 1Gabriel argues that the WSIS framework has stood the test of time, with its architecture, action lines, and multi-stakeholder processes all remaining valid and effective. She further contends that the Global Digital Compact has extended WSIS's function as an implementation engine rather than replacing it.
She noted that last December the General Assembly adopted the WSIS Plus 20 outcome resolution, and that the international community looked at what had been built and affirmed that the architecture, action lines, and multi-stakeholder processes hold . She stated that with the Global Digital Compact adopted in between, WSIS has extended its function as an implementation engine .
on: The WSIS framework has proven its value and must now shift decisively from dialogue and reflection to concrete implementation
UNESCO facilitates action lines covering access, education, cultural diversity, and media freedom as inseparable pillars of inclusive knowledge societies
Arg. 2Gabriel argues that UNESCO's facilitation of five WSIS action lines is not a matter of institutional habit but reflects the inseparable nature of access, education, sciences, cultural diversity, media freedom, and AI ethics in building truly inclusive knowledge societies. She contends that these pillars cannot be addressed in isolation.
She stated that UNESCO facilitates five WSIS action lines because access, education, sciences, cultural diversity, media freedom, and the ethics of new technologies including AI are inseparable . She argued that one cannot build inclusive knowledge societies that leave out women, speak only dominant languages, or ignore young people .
on: The digital divide — particularly between the global north and south, and between connected and unconnected populations — remains a critical and urgent challenge
UNESCO invests in digital competency frameworks for teachers everywhere and in low-resource AI tools that work without large data centres
Arg. 3Gabriel highlights UNESCO's practical investments in tools and frameworks designed specifically for contexts where resources are limited, including AI solutions that function without large data centres and digital competency frameworks tailored for teachers globally. She presents these as concrete expressions of UNESCO's commitment to inclusive digital development.
She stated that UNESCO invests in low-resource AI, in edge computing that works without data centres, and in digital competency frameworks designed for teachers everywhere . She also referenced UNESCO's work in classrooms, newsrooms, national AI policies in small island states, and youth parliaments as examples of where global ambitions become real for real people .
on: Education — particularly teacher training — is the most effective and scalable lever for digital and AI inclusion
The investment boom in AI and digital infrastructure is geographically narrow, with least developed countries capturing only 10% of strategic sector investments
Arg. 1Moreno argues that while the world is experiencing an unprecedented investment boom in AI and digital infrastructure, the benefits are highly concentrated geographically, leaving least developed and lower middle-income countries largely excluded. He contends that this narrow geography of investment is a fundamental challenge to equitable digital development.
He stated that the world will spend $800 billion on AI infrastructure in the current year alone, and that projections to 2030 may make this the biggest capital expenditure in modern history . He noted that strategic sectors like AI, semiconductors, and clean energy attract half of all new investments worldwide, but that least developed and lower middle-income countries capture only 10% of strategic sector investments .
on: The digital divide — particularly between the global north and south, and between connected and unconnected populations — remains a critical and urgent challenge
There is a critical time gap between the pace of AI development and the construction timelines for infrastructure in developing nations, risking exclusion from rule-setting
Arg. 2Moreno argues that the mismatch between the rapid pace of AI development and the much longer timelines required to build physical and educational infrastructure in developing countries means that most of the world will arrive too late to shape the rules of the digital economy. He contends that global cooperation is the only mechanism that can bridge this time gap.
He outlined the contrasting timelines: a data centre takes three years to build, a power plant five years, a grid ten years, and an education system a generation, while the AI frontier runs on quarterly cycles and is accelerating . He warned that even with the right investments and policies, building alone gets most countries to the table after the rules are already written .
UNCTAD chairs the UN Group on the Information Society and launched a task force on financing for digital development to address the investment gap facing developing countries
Arg. 3Moreno presents UNCTAD's institutional contributions to bridging the digital investment gap, including chairing the UN Group on the Information Society and launching a dedicated task force on financing for digital development. He frames these as concrete actions to align WSIS and the Global Digital Compact and address the financing challenges facing developing countries.
He stated that UNCTAD had the honour of chairing the UN Group on the Information Society, which delivered a joint implementation roadmap to align WSIS and the Global Digital Compact . He also noted the launch of a task force on financing for digital development aimed at the investment gap he described .
The WSIS framework has proven remarkably prescient over 20 years and must now shift from reflection to implementation
Arg. 1Opp argues that the WSIS Plus 20 review reaffirmed a vision that has been remarkably prescient for two decades, but that the current moment demands a decisive shift from dialogue and reflection to concrete implementation. He stresses that the next decade must be focused on delivering results together.
He stated that the WSIS Plus 20 review reaffirmed the WSIS vision of an inclusive, people-centred, and development-oriented digital future . He explicitly stated that today has to be more than reflection and must be about implementation, with the next decade focused on delivering results together .
on: The WSIS framework has proven its value and must now shift decisively from dialogue and reflection to concrete implementation
Multi-stakeholder partnership is not merely support for implementation; it is the implementation model itself, requiring government, UN, private sector, civil society, and youth
Arg. 2Opp argues that multi-stakeholder partnership is not simply a supporting mechanism for implementation but is the implementation model itself, as no single institution possesses all the capabilities needed to deliver on the WSIS vision. He presents this as the bedrock of the WSIS process.
He stated that multi-stakeholder partnership is not simply support for implementation but is the implementation model itself, requiring government, the UN system, the private sector, academia, civil society, the technical community, and young innovators, each bringing something no single institution can possess .
on: Multi-stakeholder partnership — encompassing governments, private sector, civil society, academia, and the technical community — is essential for implementing the digital future
UNDP's work across 26 countries shows the key question is not whether countries are AI-ready, but whether they are building the capability to steer AI toward public value
Arg. 3Opp argues that the framing of AI readiness is insufficient and that the more important question is whether countries are developing the institutional capability to direct AI towards public value. He presents UNDP's research and country-level work as evidence for this reframing.
He referenced a UNDP report based on work across 26 countries on AI transformation, which found that the question is not whether countries are AI-ready but whether they are building the capability to steer AI toward public value . He cited concrete examples from Zimbabwe and Ghana, where UNDP's support for national AI strategies was paired with training for policymakers to move from strategy to implementation , and from Trinidad and Tobago, where public procurement was used as a governance lever for managing AI adoption .
on: The digital divide — particularly between the global north and south, and between connected and unconnected populations — remains a critical and urgent challenge
The ITU Secretary General framed the forum's central challenge as leadership that inspires action, builds trust, and delivers results for a more inclusive digital world
Arg. 1Bogdan-Martin opens the forum by framing its central question as one of leadership — specifically, the kind of leadership that can inspire action, build trust, and deliver tangible results for a more inclusive digital world. She argues that this people-centred approach to technology must remain at the centre of everything the international community does.
She stated that the question before the forum is one of leadership that inspires action, builds trust, and delivers results for a more inclusive digital world, and that it is a determining factor of the future with stakes that could not be higher . She noted that as digital transformation accelerates, many divides persist and risk widening, making a people-centred approach essential .
on: Digital inclusion requires more than connectivity; meaningful access must include digital skills, literacy, and the ability to use technology productively
Universal meaningful connectivity requires that no one is left behind, and this must be the central measure of true digital inclusion
Arg. 1Abdullah frames the core challenge of the forum as achieving universal meaningful connectivity in a way that ensures no person is excluded. She presents this not merely as a technical goal but as a standard of genuine digital inclusion that all panellists should address.
In posing the rapid-fire question to all panellists, she asked what single change they would make to accelerate universal meaningful connectivity and ensure true digital inclusion, explicitly defining this as meaning no one is left behind .
The WSIS Forum must transition from dialogue to delivery, moving from the information society to the intelligence society
Arg. 2Abdullah argues that the forum's people-centred vision must serve as the reference point as the international community moves beyond discussion towards concrete implementation. She frames this shift as a progression from the information society to what the ITU Secretary General has called the intelligence society.
In her closing remarks, she stated that the people-centred WSIS vision remains the reference point and invoked the Secretary General's framing of the transition from the information society to the intelligence society, emphasising the move from dialogue to delivery .
on: The WSIS framework has proven its value and must now shift decisively from dialogue and reflection to concrete implementation
The co-organiser model of ITU, UNESCO, UNDP, and UNCTAD reflects the multi-stakeholder and multi-agency partnership that underpins the WSIS Forum
Arg. 3Abdullah highlights that the WSIS Forum is jointly organised by four major UN entities — ITU, UNESCO, UNDP, and UNCTAD — and that this co-organiser structure has been in place for many years. She presents this institutional partnership as a defining feature of the forum's legitimacy and reach.
She introduced the closing segment by explicitly naming ITU, UNESCO, UNDP, and UNCTAD as co-organisers of the forum, noting that they have been co-organisers for many years, and invited representatives of each to take the stage .
on: Multi-stakeholder partnership — encompassing governments, private sector, civil society, academia, and the technical community — is essential for implementing the digital future
Session Knowledge Graph
Speakers · Topics · Arguments · Relationships
Karis noted that the central question of how digital transformation serves people remains the same as it was over 20 years ago , while Gabriel affirmed that the WSIS architecture, action lines, and multi-stakeholder processes all hold and have been extended by the Global Digital Compact . Opp explicitly stated that today must be more than reflection and must be about implementation, with the next decade focused on delivering results together . Abdullah echoed this in her closing remarks, invoking the transition from the information society to the intelligence society and the move from dialogue to delivery .
The WSIS framework has proven remarkably prescient over 20 years and must now shift from reflection to implementation
The WSIS architecture, action lines, and multi-stakeholder processes remain valid and have been extended by the Global Digital Compact
The WSIS framework has proven remarkably prescient over 20 years and must now shift from reflection to implementation
The WSIS Forum must transition from dialogue to delivery, moving from the information society to the intelligence society
Karis described Estonia's 'tiger leap' of providing internet access to all schools and its current 'AI leap', with 100% of teachers and students now involved in AI education . Madiyev reported that Kazakhstan covered one million people with AI skills in a single year and is launching an AI University . Battsengel shared direct evidence that training 1,000 teachers from 330 villages in Mongolia led to those teachers training 50,000 children , arguing that teachers are the multipliers in the education system . Gabriel highlighted UNESCO's investment in digital competency frameworks designed for teachers everywhere .
Education is the single most important factor for digital transformation; Estonia prioritised internet access in all schools and is now integrating AI literacy for all teachers and students
Kazakhstan covered one million people with AI skills in one year and is launching an AI University, reflecting the scale of investment needed in human capital
Training teachers as multipliers is more effective than focusing solely on students; 1,000 trained teachers in Mongolia reached 50,000 children in rural communities
UNESCO invests in digital competency frameworks for teachers everywhere and in low-resource AI tools that work without large data centres
Baerbock argued that advanced AI governance frameworks are futile if large portions of the population lack basic internet access or electricity . Battsengel explicitly stated that digital inclusion does not mean using social media like Instagram or TikTok, but using connectivity to get more opportunities to learn and embrace technology , and that the last-mile child needs not only connectivity but access to digital platforms to learn and create . Bogdan-Martin thanked Battsengel for stressing the importance of not just access but the skilling piece to make access meaningful . Gabriel argued that access, education, cultural diversity, and AI ethics are inseparable pillars that cannot be addressed in isolation .
Billions remain unconnected, and without basic electricity and internet access, advanced technologies like AI are meaningless for the most vulnerable
Digital inclusion must go beyond connectivity to ensure meaningful access, including digital skills and learning platforms for last-mile communities
The ITU Secretary General framed the forum's central challenge as leadership that inspires action, builds trust, and delivers results for a more inclusive digital world
UNESCO facilitates action lines covering access, education, cultural diversity, and media freedom as inseparable pillars of inclusive knowledge societies
Karis stated that AI must remain human-centred, rights-based, and accountable, and that digital transformation can strengthen societies only when it allows for civic space and democratic accountability . Satyarthi argued that technology carries the minds, purposes, intentions, and values of those who create it , and asked whether AI systems will protect the last child of society or exploit them further . Battsengel stated that blocking access to digital skills violates human rights to learn and receive equal education . Baerbock argued that without liability for AI, human rights and human dignity are endangered .
Human rights must be protected online just as offline, and AI must remain human-centred, rights-based, and accountable
Technology is not neutral; it carries the values and intentions of its creators, and leaders must ensure AI serves and protects the most vulnerable, including children
Denying children and last-mile communities access to digital education and AI tools is a violation of their human rights to learn and develop equally
Liability for big tech companies must be established to protect human dignity, just as liability was established for other powerful technologies such as nuclear energy
Karis argued that governments cannot govern the digital world alone and that the private sector, civil society, the technical community, and academia all have a central role to play . Opp stated that multi-stakeholder partnership is not simply support for implementation but is the implementation model itself, as no single institution possesses all the capabilities needed . Battsengel called for tech and AI companies to be brought to the table on AI education, human rights, and inclusion discussions . Gabriel affirmed UNESCO's commitment to multi-stakeholder processes as the foundation of inclusive knowledge societies . Abdullah highlighted the co-organiser model of ITU, UNESCO, UNDP, and UNCTAD as a defining feature of the forum .
Digital governance must be inclusive of all stakeholders — governments, private sector, civil society, academia, and the technical community
Multi-stakeholder partnership is not merely support for implementation; it is the implementation model itself, requiring government, UN, private sector, civil society, and youth
Tech companies and AI firms must be brought to the table on AI education, inclusion, and human rights discussions, as many are currently absent from these conversations
UNESCO facilitates action lines covering access, education, cultural diversity, and media freedom as inseparable pillars of inclusive knowledge societies
The co-organiser model of ITU, UNESCO, UNDP, and UNCTAD reflects the multi-stakeholder and multi-agency partnership that underpins the WSIS Forum
Baerbock highlighted that millions in sub-Saharan Africa are not connected to the internet or even to electricity, making AI solutions irrelevant without basic infrastructure . Moreno reported that least developed and lower middle-income countries capture only 10% of strategic sector investments, with the world spending $800 billion on AI infrastructure in the current year alone . Battsengel referenced ITU reports indicating that 2.2 billion people have not been able to access digital skills or the internet . Gabriel noted that billions cannot access the internet, and millions of teachers and learners are navigating digitalisation with tools not built for their languages or realities . Opp acknowledged that the digital divide and all the things that need to be done remain as gaps .
Billions remain unconnected, and without basic electricity and internet access, advanced technologies like AI are meaningless for the most vulnerable
The investment boom in AI and digital infrastructure is geographically narrow, with least developed countries capturing only 10% of strategic sector investments
Digital inclusion must go beyond connectivity to ensure meaningful access, including digital skills and learning platforms for last-mile communities
UNESCO facilitates action lines covering access, education, cultural diversity, and media freedom as inseparable pillars of inclusive knowledge societies
UNDP's work across 26 countries shows the key question is not whether countries are AI-ready, but whether they are building the capability to steer AI toward public value
Both Karis and Battsengel independently arrived at the same strategic conclusion: that training teachers is the most effective way to scale digital and AI education. Karis described how Estonia started to teach teachers as part of its AI leap, with 100% of teachers involved in the second year of the programme . Battsengel shared that from her direct experience, one of the key lessons for government leaders is to start with teachers because teachers are the multipliers in the education system , and that 1,000 trained teachers subsequently trained 50,000 children across 330 villages in Mongolia . Both Satyarthi and Battsengel framed the exclusion of children and marginalised communities from digital and AI systems as a fundamental moral and human rights issue. Satyarthi asked whether AI systems will protect the last child of society or exploit them further , and called for compassion to be embedded in AI algorithms . Battsengel stated that blocking access to digital skills is not only about connectivity but violates human rights to learn and receive equal education , and emphasised that it is the responsibility of all to build AI for every child in the last-mile zones . Both Baerbock and Battsengel explicitly framed internet access and digital inclusion as matters of human rights rather than merely development goals. Baerbock called for a formal right to internet access, arguing that without it, other human rights including housing and education are endangered . Battsengel stated that the 2.2 billion people without access to digital skills and internet are being blocked in a way that violates their human rights to learn and receive equal education . Moreno, Baerbock, and Gabriel all highlighted the geographic and structural concentration of digital investment and its consequences for the global south. Moreno provided the sharpest quantitative framing, noting that least developed and lower middle-income countries capture only 10% of strategic sector investments despite the world spending $800 billion on AI infrastructure in the current year . Baerbock argued that millions in sub-Saharan Africa are not connected to the internet or even to electricity, making AI solutions irrelevant . Gabriel noted that AI sits on a foundation that too many people in the global south or in rural areas cannot rely upon . Both Karis and Madiyev, representing countries at different stages of digital development, shared a strong emphasis on large-scale, government-led investment in AI education and human capital. Karis described Estonia's systematic approach of providing internet access to all schools and then integrating AI education for all teachers and students . Madiyev reported that Kazakhstan covered one million people with AI skills in a single year, is launching an AI University, and has established the ALAM AI International AI Centre to cover broad population groups . Madiyev also explicitly thanked Karis for sharing Estonia's experiences . All four speakers affirmed the enduring validity of the WSIS framework and the importance of ensuring that the Global Digital Compact and other initiatives reinforce rather than duplicate it. Baerbock described the WSIS Plus 20 consensus as a rare and powerful moment that should be leveraged . Karis explicitly stated as his fifth principle that the WSIS process and other initiatives should reinforce each other and not create duplication . Opp stated that the WSIS Plus 20 review reaffirmed a vision that has been remarkably prescient for 20 years . Gabriel affirmed that the WSIS architecture holds and that the Global Digital Compact has extended its function as an implementation engine .
It was unexpected that a Nobel Peace laureate known for child labour activism and a tech entrepreneur running an AI academy would converge so directly on the same practical proposal. Satyarthi called on industry to give their engineers one month to meet children in mines, factories, war zones, and remote villages, arguing that this experience would fundamentally change how they design AI systems . Battsengel responded directly to this, saying she would be happy to walk with him and would give her own engineers for one month as well . This convergence between a moral activist and a technology practitioner on a concrete, actionable proposal for humanising AI development was a notably unexpected moment of consensus.
It was somewhat unexpected that a head of state and president of the UN General Assembly, a Central Asian tech entrepreneur, and a Nobel Peace laureate from India all converged on framing digital access explicitly as a human rights issue. Baerbock acknowledged this was controversial but called for a formal right to internet access , arguing that its absence endangers rights to housing and education . Battsengel stated that blocking access to digital skills violates human rights to learn and receive equal education . Satyarthi, while approaching from the angle of child protection, framed the question of who AI systems serve and protect as a fundamental moral and rights-based question . The breadth of this consensus across very different professional and geographic backgrounds was notable.
It was unexpected that the Deputy Prime Minister of Kazakhstan, a European political leader, and a Mongolian tech entrepreneur would all converge on the idea that AI must be treated as foundational public infrastructure. Madiyev explicitly proposed that AI should be the base infrastructure for people going forward, like water supply, electricity, and internet . Baerbock argued that without basic electricity and internet, AI solutions are meaningless for the most vulnerable , implicitly treating connectivity as a prerequisite infrastructure. Battsengel argued that last-mile children need not only connectivity but access to digital platforms and AI tools as a matter of right . The convergence of a government minister, a political leader, and a civil society entrepreneur on this framing was a notable area of unexpected agreement.
The discussion revealed a remarkably high level of consensus across all speakers on several core themes: the enduring validity and need to implement the WSIS framework; the centrality of education and teacher training for digital and AI inclusion; the insufficiency of connectivity alone without skills and meaningful access; the imperative that AI be human-centred, rights-based, and serve the most vulnerable; the necessity of multi-stakeholder partnership as the implementation model itself; and the urgency of addressing the persistent and potentially widening digital divide between the global north and south. Speakers from very different backgrounds — heads of state, a Nobel laureate, tech entrepreneurs, and UN officials — converged on these themes with striking consistency. There were also unexpected areas of consensus, particularly around framing internet access as a human right, treating AI as essential public infrastructure, and the proposal for engineers to engage directly with marginalised communities.
When asked what single change they would make to accelerate universal meaningful connectivity , speakers gave markedly different answers reflecting divergent priorities. Baerbock called for a formal right to internet access, arguing that without it, other human rights such as housing and education are endangered . Karis consistently returned to education as the cornerstone, citing Estonia's 'tiger leap' and current AI literacy programme covering 100% of teachers and students . Madiyev proposed elevating AI to the status of base infrastructure like water and electricity . Satyarthi argued the most important change was to alter who builds AI systems, by sending engineers to spend a month with marginalised children . Battsengel focused on ensuring that connectivity is accompanied by meaningful digital learning platforms, not merely social media access .
A right to internet access should be formally guaranteed, as its absence endangers other fundamental human rights including housing and education
Education is the single most important factor for digital transformation; Estonia prioritised internet access in all schools and is now integrating AI literacy for all teachers and students
AI should be considered base infrastructure for people, equivalent to water supply and electricity, to prevent widening inequality
Engineers and AI developers should spend time with marginalised children to understand the human consequences of the code they write, fostering empathy-driven design
Digital inclusion must go beyond connectivity to ensure meaningful access, including digital skills and learning platforms for last-mile communities
Baerbock placed strong emphasis on formal legal mechanisms, calling for a guaranteed right to internet access and liability frameworks for big tech companies analogous to those established for nuclear technology . Madiyev also favoured legislative approaches, having embedded digital rights in Kazakhstan's constitution and passed a suite of digital laws , but his practical emphasis was on building infrastructure and human capital. Karis and Battsengel, by contrast, focused primarily on education and skills as the most effective levers , with less emphasis on formal rights frameworks. This reflects a tension between a rights-based legal approach and a capacity-building practical approach to achieving the same goal of digital inclusion.
A right to internet access should be formally guaranteed, as its absence endangers other fundamental human rights including housing and education
Liability for big tech companies must be established to protect human dignity, just as liability was established for other powerful technologies such as nuclear energy
Education is the single most important factor for digital transformation; Estonia prioritised internet access in all schools and is now integrating AI literacy for all teachers and students
Kazakhstan has embedded digital rights and personal data protection directly into its constitution, reflecting a comprehensive legislative approach to digitalisation
Building AI skills is no longer just about training; it is about building economic engines, creating competitiveness, and filling opportunity gaps
Madiyev framed AI in largely techno-economic terms, arguing it should be treated as base infrastructure like water supply and electricity to deliver productivity gains and prevent inequality . He emphasised building computing clusters, AI laws, and entrepreneurial ecosystems . Satyarthi, by contrast, argued that technology is not neutral and carries the values and intentions of its creators , and that the fundamental question is not what AI can do but who it will serve and protect . He called for compassion - encompassing justice, equality, peace, and sustainability - to be embedded in AI algorithms as a disruptive and transformative force . These represent genuinely different conceptual framings of what AI is and what the primary challenge of governing it entails.
AI should be considered base infrastructure for people, equivalent to water supply and electricity, to prevent widening inequality
Compassion must be embedded as a transformative force — encompassing justice, equality, peace, and sustainability — in the development and deployment of AI algorithms
Technology is not neutral; it carries the values and intentions of its creators, and leaders must ensure AI serves and protects the most vulnerable, including children
Baerbock made an explicit and forceful call for liability frameworks for big tech companies, drawing a direct analogy with nuclear technology and arguing that without such accountability, human rights and human dignity are endangered . No other panellist endorsed or even addressed this specific proposal. Karis framed the private sector as a necessary partner in inclusive multi-stakeholder governance but did not call for liability. Battsengel called for tech companies to be brought into AI education and inclusion discussions , framing the issue as one of engagement rather than legal accountability. Madiyev and Satyarthi did not address the liability question at all. This represents a significant divergence in how Baerbock conceptualises the relationship between states and the private sector compared to her fellow panellists.
Liability for big tech companies must be established to protect human dignity, just as liability was established for other powerful technologies such as nuclear energy
Digital governance must be inclusive of all stakeholders — governments, private sector, civil society, academia, and the technical community
Tech companies and AI firms must be brought to the table on AI education, inclusion, and human rights discussions, as many are currently absent from these conversations
In a forum broadly united around inclusive digital development, an unexpected divergence emerged around who should be building AI systems and what motivates them. Satyarthi made the striking proposal that industry should send engineers to spend one month with marginalised children in mines, factories, and war zones, arguing this experiential empathy would fundamentally change how they design AI . This implicitly critiques the current composition and values of the AI development community. By contrast, Madiyev's vision was of an entrepreneurial youth culture motivated by building unicorn companies , and Karis focused on broad AI literacy for all students . These represent genuinely different visions of who the ideal AI developer is and what should motivate them - empathy and moral responsibility versus entrepreneurial ambition versus broad civic literacy - which was not a tension explicitly flagged by the moderators but emerged from the juxtaposition of their responses.
In a forum that consistently celebrated multi-stakeholder partnership, an unexpected tension emerged around the nature of the relationship with the private sector. Baerbock called for legal liability for big tech companies , framing them as powerful actors requiring regulatory constraint. Battsengel noted that many tech companies are currently absent from AI education and inclusion discussions and need to be brought to the table , suggesting a more collaborative but currently unfulfilled partnership. Karis and Opp framed the private sector as an equal and essential partner in multi-stakeholder governance . This divergence - between the private sector as a partner, an absent actor needing engagement, or a powerful entity requiring legal accountability - was unexpected given the forum's generally collaborative tone and was not directly debated.
Baerbock unexpectedly raised the peace and security dimension of digital infrastructure, specifically citing the vulnerability of undersea internet cables to attack by container ships and the absence of regulatory protection . She argued that the three pillars of the UN - development, human rights, and peace and security - are deeply interconnected and must all be addressed in digital governance . This dimension was not raised by any other panellist. Karis touched on security in the context of an open and secure internet but did not address physical infrastructure security or geopolitical threats. Other panellists focused entirely on development and human rights dimensions. This created an implicit disagreement about the scope of what digital governance forums like WSIS should address.
The discussion was characterised by a high degree of surface-level consensus around shared goals - inclusive digital development, bridging the digital divide, people-centred AI, and the continued relevance of the WSIS framework - but revealed meaningful divergences in priorities, mechanisms, and conceptual framings when speakers were pressed for specific recommendations. The most substantive disagreements concerned: (1) what single intervention would most accelerate digital inclusion, with answers ranging from a legal right to internet access to education to treating AI as base infrastructure to empathy-driven developer culture ; (2) whether legal accountability mechanisms for big tech are necessary or whether partnership and engagement are sufficient ; (3) whether AI is primarily an infrastructure and economic challenge or a moral and ethical one ; and (4) whether peace and security should be integrated into digital governance discussions . There were also partial agreements where speakers shared goals but differed on the most effective pathway, particularly around education strategies and the interpretation of WSIS coherence.
All four speakers agreed that education and skills development are essential for digital inclusion and AI readiness . However, they differed on the most effective strategy within education. Karis emphasised a national top-down approach covering 100% of teachers and students . Madiyev focused on scale and breadth, covering one million people with AI skills and launching an AI University . Battsengel argued specifically that training teachers as multipliers is more effective than focusing on students directly, citing the example of 1,000 teachers reaching 50,000 children . Gabriel emphasised the need for tools designed for teachers everywhere, including low-resource AI and edge computing . They share the goal of AI education but differ on the most effective entry point and methodology.
Education is the single most important factor for digital transformation; Estonia prioritised internet access in all schools and is now integrating AI literacy for all teachers and students Kazakhstan covered one million people with AI skills in one year and is launching an AI University, reflecting the scale of investment needed in human capital Training teachers as multipliers is more effective than focusing solely on students; 1,000 trained teachers in Mongolia reached 50,000 children in rural communities UNESCO invests in digital competency frameworks for teachers everywhere and in low-resource AI tools that work without large data centres
All four speakers agreed that the digital divide remains a critical and urgent problem . However, they differed in their diagnosis and proposed solutions. Baerbock focused on the absence of basic electricity and internet as the foundational barrier , and proposed a formal right to internet access . Battsengel argued that connectivity alone is insufficient and must be accompanied by meaningful digital learning platforms . Moreno framed the problem in terms of investment geography and the time gap between AI development and infrastructure construction in developing countries . Karis addressed the inclusion gap through the lens of digital literacy and the need for AI systems to work in local languages and contexts . They agree on the problem but propose different primary interventions.
Billions remain unconnected, and without basic electricity and internet access, advanced technologies like AI are meaningless for the most vulnerable Digital development must be inclusive Digital inclusion must go beyond connectivity to ensure meaningful access, including digital skills and learning platforms for last-mile communities The investment boom in AI and digital infrastructure is geographically narrow, with least developed countries capturing only 10% of strategic sector investments
All four speakers agreed on the value and continued relevance of the WSIS framework and the need for coherence across global digital governance processes . However, they emphasised different aspects of what coherence requires. Karis focused on avoiding duplication between WSIS and other initiatives . Baerbock stressed the need to connect WSIS to other UN processes, including climate commitments and peace and security . Opp emphasised the shift from reflection to implementation as the defining challenge . Gabriel highlighted that WSIS was always about people rather than technology and that this vision must guide implementation . They agree the framework is valid but differ in what they see as the primary gap in its implementation.
Global initiatives such as WSIS and the Global Digital Compact must reinforce each other rather than create duplication The three pillars of the United Nations — development, human rights, and peace and security — are deeply interconnected and must be addressed together in digital governance The WSIS framework has proven remarkably prescient over 20 years and must now shift from reflection to implementation The WSIS architecture, action lines, and multi-stakeholder processes remain valid and have been extended by the Global Digital Compact
Baerbock, Satyarthi, and Battsengel all agreed that the most vulnerable — particularly children and those in last-mile communities — must be at the centre of digital governance decisions . However, they differed significantly on how to achieve this. Baerbock focused on structural and legal mechanisms such as a right to internet access and big tech liability . Satyarthi called for a cultural and empathetic transformation among AI developers, proposing that engineers spend a month with marginalised children . Battsengel focused on practical access to digital learning platforms and the role of teachers as multipliers . They share a people-centred and child-centred vision but diverge on the mechanisms for realising it.
Billions remain unconnected, and without basic electricity and internet access, advanced technologies like AI are meaningless for the most vulnerable Technology is not neutral; it carries the values and intentions of its creators, and leaders must ensure AI serves and protects the most vulnerable, including children Denying children and last-mile communities access to digital education and AI tools is a violation of their human rights to learn and develop equally
- Multilateralism is under pressure but remains alive, as demonstrated by the consensus adoption of the WSIS Plus 20 resolution — a rare moment of global agreement that must be built upon rather than taken for granted.
- The WSIS framework, its action lines, and multi-stakeholder processes remain valid and relevant after 20 years; the priority must now shift decisively from reflection and dialogue to concrete implementation and delivery of results.
- Billions of people remain unconnected, and without foundational infrastructure — electricity, internet access, and digital skills — advanced technologies such as AI are meaningless for the most vulnerable populations, particularly in sub-Saharan Africa and rural communities globally.
- A formal right to internet access should be established, as its absence directly endangers other fundamental human rights including education, housing, and economic participation.
- The global investment boom in AI and digital infrastructure is geographically concentrated, with least developed countries capturing only 10% of strategic sector investments, creating a dangerous and widening divide.
- There is a critical mismatch between the rapid pace of AI development and the long construction timelines for infrastructure in developing nations, meaning many countries risk being excluded from shaping the rules of the digital economy.
- Digital inclusion must extend beyond mere connectivity to encompass meaningful access — including digital literacy, AI skills, and access to learning platforms — particularly for last-mile and marginalised communities.
- Education is the single most critical lever for digital transformation; training teachers as multipliers is a highly effective strategy, as demonstrated by Mongolia where 1,000 trained teachers reached 50,000 rural children.
- AI must be treated as base infrastructure equivalent to water and electricity, and must be governed by principles of openness, human rights protection, inclusivity, and accountability.
- Technology is not neutral; it carries the values and intentions of its creators, and compassion — encompassing justice, equality, peace, and sustainability — must be embedded in the development and deployment of AI systems.
- Liability for big tech companies must be formally established to protect human dignity, drawing on precedents set for other powerful technologies such as nuclear energy.
- Multi-stakeholder partnership — encompassing governments, the UN system, the private sector, civil society, academia, and youth — is not merely a support mechanism for implementation but is the implementation model itself.
- The three pillars of the United Nations — development, human rights, and peace and security — are deeply interconnected and must be addressed in an integrated manner within digital governance frameworks.
- Global initiatives such as WSIS and the Global Digital Compact must reinforce each other and avoid duplication, working as a coherent system rather than competing frameworks.
- Cross-regional leadership alliances are needed to connect WSIS action lines and ensure that those still excluded from digital progress are not left further behind.
“Technology is not neutral. Helicopters are not neutral. They carry the minds and purpose and intentions and values of those people who create them. So this is the time to ensure that right from the development and deployment, the people who are involved in it have to change... compassion should flow in your brain and your veins but also compassion should flow — the fire of compassion should flow in the codes and fibers.”
“If we are not connecting these to the other UN processes — for example, delivering on the civilian commitments, delivering on the commitments to bridging divides also financially with technology... if still a fifth, if not to say a fourth, of the population doesn't even have internet access, it doesn't help at all to have an AI model which might deliver on e-health if the basic services are not there.”
“My one change is to change who is sitting in this room building AI systems and writing codes, because most of you have never met a child who could be failed one day due to the codes you write... Give your engineers for one month. I invite them to meet those children in mines and factories, children who are living in war zones, always in danger, always in trauma.”
“The world will spend $800 billion on AI infrastructure this year alone... But the geography of this boom is narrow. Two economies host most of it. Least developed and lower middle-income countries capture only 10% of strategic sector investments. Behind the investment divide sits a divide of time. A data centre takes three years to build. A power plant takes five. A grid takes ten. An education system takes a generation. Infrastructure runs on construction time. The frontier of AI and the digital economy runs on quarters.”
“First thing would be to guarantee a right to internet access... And if this internet access is being misused, the question of liability of big tech companies is the second point we really have to nail down. We did this with all technologies in all fields before with revolutionary changes, especially with nuclear. We guaranteed that there's a liability of those who hold this power in their hands.”
“We need people who have skills... Building AI skills no longer means training. It means building an economic engine. It means creating competitiveness. It means creating opportunities and filling gaps... 1,000 teachers trained from 330 villages in Mongolia trained 50,000 kids — kids coming from riding horses, getting raised in the wilderness — to build their AI agents.”
“Digital governance must remain inclusive of all stakeholders. Governments have a crucial role in setting rules and protecting rights, but they cannot govern the digital world alone... Human rights must be protected online just as they are protected offline... For languages, inclusion also has a cultural dimension. The system must work not only in the largest global languages, but also understand local languages, cultures, and contexts.”
How can national digital economies be better connected to global frameworks to ensure inclusive growth, particularly for those still lacking basic electricity and internet access?
Baerbock highlighted that even well-designed AI governance frameworks are ineffective if large portions of the population—particularly in sub-Saharan Africa—lack electricity and internet access. This raises the need for further research into how WSIS action lines and SDG commitments can be better integrated to address foundational infrastructure gaps alongside digital governance.
How can undersea cable infrastructure be better protected through international regulation, given the growing risk of deliberate sabotage?
Baerbock raised the specific concern that undersea cables—critical to global internet connectivity—currently lack adequate international protection or regulation. This is an emerging security and governance gap that warrants further research and multilateral policy development.
Should internet access be enshrined as a formal human right, and what legal and institutional frameworks would be needed to implement this?
Baerbock explicitly called for a 'right to internet access,' noting that the absence of connectivity is already undermining other established human rights such as housing, education, and banking. This is a contested but important area requiring further legal, policy, and normative research.
How should liability for big tech companies be structured in the context of AI, and what precedents from other transformative technologies (e.g., nuclear) could inform this?
Baerbock drew a parallel between AI and nuclear technology, arguing that liability frameworks must be established for those who hold power over AI systems. This is an underexplored area of AI governance that requires further legal and policy research.
How can AI systems be designed to function effectively in low-resource environments, including support for local languages, cultures, and contexts beyond dominant global languages?
President Karis noted that digital inclusion has a cultural dimension and that AI systems must work not only in major global languages but also in local languages and contexts. This points to a significant research gap in multilingual and culturally adaptive AI development.
How can global digital initiatives such as the Global Digital Compact, WSIS, and national programmes be better coordinated to avoid duplication and reinforce each other?
Karis emphasised that international digital initiatives should reinforce rather than duplicate each other. Further research is needed into governance architectures and coordination mechanisms that align these overlapping frameworks effectively.
How can AI be treated as foundational public infrastructure—akin to water, electricity, and the internet—and what policy frameworks would support this?
Deputy Prime Minister Madiyev suggested that AI should be considered basic infrastructure for all people, not just a technological tool. This raises important questions about public investment models, universal access policies, and regulatory frameworks that would need further investigation.
How can entrepreneurial and innovation ecosystems be embedded within university education systems to foster a generation of technology founders, particularly in developing economies?
Madiyev proposed merging startup ecosystems with universities so students can receive academic credit for incubation and acceleration programmes. This is an area requiring further research into educational policy, economic development, and innovation ecosystem design.
Who is currently involved in building AI systems and writing code, and how can the diversity of those developers be broadened to include perspectives from the most vulnerable communities?
Satyarthi argued that AI developers are largely disconnected from the children and communities most affected by their systems. This raises important questions about diversity, inclusion, and ethical responsibility in the AI workforce, warranting further research into recruitment, training, and exposure programmes.
How can compassion be systematically embedded into AI development and deployment processes, and what governance or design mechanisms would support this?
Satyarthi called for compassion—defined as encompassing justice, equality, peace, and sustainability—to be embedded in AI algorithms and digital systems. This is a novel and underexplored area at the intersection of AI ethics, design philosophy, and human rights that merits further research.
How can AI skilling programmes be designed to prioritise teachers as multipliers, and what models have proven most effective in reaching last-mile and rural communities?
Battsengel shared evidence that training teachers first produced a multiplier effect, enabling 1,000 trained teachers to reach 50,000 children across 330 villages. Further research is needed to understand how this model can be scaled, adapted, and evaluated across different national and cultural contexts.
How can technology companies be more effectively brought into AI education, inclusion, and human rights discussions, and what collaborative models would be most impactful?
Battsengel noted that many AI and tech companies are not participating in AI education and inclusion discussions. Research is needed into incentive structures, partnership models, and accountability mechanisms that could bring industry more meaningfully into these conversations.
How can digital inclusion be defined and measured beyond mere connectivity to encompass meaningful access, skills, and the ability to create economic opportunities?
Battsengel stressed that digital inclusion is not simply about access to social media but about enabling people to learn, create, and build opportunities. This calls for further research into metrics, frameworks, and policy definitions that capture the full spectrum of meaningful digital participation.
How can the significant investment gap in AI and digital infrastructure between developed economies and least developed countries be addressed through global cooperation and financing mechanisms?
Moreno highlighted that $800 billion will be spent on AI infrastructure globally in 2025, yet least developed and lower middle-income countries capture only 10% of strategic sector investments. This stark disparity requires urgent research into financing mechanisms, investment policy, and international cooperation frameworks.
How can global cooperation bridge the 'time gap' between the slow pace of infrastructure development and the rapid acceleration of AI and the digital economy, so that developing countries can shape the rules before they are written?
Moreno articulated a structural problem: infrastructure takes years or decades to build, while AI advances on quarterly cycles, meaning most countries arrive too late to influence governance decisions. Research is needed into fast-track mechanisms, interim governance arrangements, and capacity-building approaches that give developing nations a meaningful voice.
How can countries move beyond AI readiness assessments to build the capability to steer AI toward public value, and what governance models support this transition?
Opp, citing UNDP research across 26 countries, reframed the question from 'AI readiness' to 'capability to steer AI toward public value.' This distinction has significant implications for how international organisations and governments design AI governance support programmes, warranting further research and case study analysis.
How can public procurement be used as a practical governance lever for managing AI adoption in government, particularly in contexts where formal legislation has not yet been enacted?
Opp cited the example of Trinidad and Tobago, where public procurement was used as a governance tool for AI adoption ahead of formal legislation. This is an underexplored but potentially highly practical area of AI governance that merits further comparative research across different national contexts.
How can the energy generated by high-level forums such as WSIS be translated into practical, on-the-ground implementation, and what accountability mechanisms would support this?
Opp stressed that the forum must move beyond dialogue to delivery, raising the persistent challenge of translating multilateral commitments into tangible outcomes. Further research is needed into implementation models, monitoring frameworks, and accountability structures that can bridge this gap.
How can AI education be designed to ensure that low-resource AI tools and edge computing solutions reach communities without reliable data centre access, particularly in the Global South?
Gabriel highlighted UNESCO's investment in low-resource AI and edge computing that works without data centres, pointing to a significant technical and policy research gap in how AI tools can be made functional and accessible in infrastructure-poor environments.
How can digital competency frameworks for teachers be designed to be universally applicable yet sensitive to local languages, cultures, and educational realities?
Gabriel noted that millions of teachers and learners are navigating digitalisation with tools not built for their languages or realities. Research is needed into how digital and AI competency frameworks can be contextualised and localised without losing coherence or comparability across systems.
