Oxfam

• As a supporter of the Global Digital Compact which, though imperfect, represents an important step forward for global governance of digital technologies and AI, Oxfam's hope is that this first Global Dialogue on AI Governance will center the ambitions of the Compact as agreed, and call upon its signatories to do the same. Further, we believe a successful Global Dialogue will affirm – and perhaps improve upon – the aims of the Compact and will strengthen the commitments of signatories, in partnership with civil society, to pursuing its ambitions aggressively, cooperatively, and accountably. Time may not be on our side. We must collectively commit to assertive actions that can harness the rising tide of global concern and fear that AI – driven by the ambitions of the private sector, largely unaccountable to civil society – is moving too fast and without sufficient human rights due diligence. Of notable concern are the concentration of decision-making power in the hands of a few powerful technology billionaires and the political processes under their influence
• rapid data center proliferation, which threatens natural resources, the environment, public health, and affordable energy
• increasing automation of conflict and the life & death decisions made by military AI
• and the increasing automation of many forms of work, leading to growth in unemployment in many jurisdictions where social safety nets are weak or non-existent. Therefore, beyond affirming the Global Digital Compact, we seek clearly articulated and achievable commitments and demands of both member countries and their respective private sectors and civil societies that will, in a timely and coordinated way, begin to align the trajectory of AI development with the goal of a human-rights-respecting digital ecosystem.

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

Oxfam's core mission is to end inequality and the injustice of poverty at a global level while working closely with partners and communities in very differentiated contexts in over 70 countries. We view these realities as systemic outcomes, not inevitabilities specific to any one community, country, or region, or attributable to any single cause. The same is true for AI and other technologies, which are likewise part of the same intimately-linked global system, and risk perpetuating these same inequalities. Protection and promotion of human rights in a digital age is paramount - our first priority - and from this we derive all others. For the makers and propagators of technology to be held accountable to the communities impacted by them, transparency, accountability, and human oversight are essential. Likewise, achieving safe, secure, ethical, and trustworthy AI (or any other digital technology), must necessarily include accounting for and mitigating the neocolonial impacts of these technologies on socioeconomic, cultural, linguistic, and other dimensions of society in rights-respecting ways. While capacity building, open source alternatives, and interoperability of governance approaches are important, for us they are less fundamental to the pursuit of a rights-respecting digital ecosystem.

Intimately tied to the issues of human rights, inequality, and poverty is the issue of meaningful and dignified employment - or social programs designed to compensate for the loss of such work - in the age of AI. We are witnessing in some countries the early signs of unemployment and displacement of workers due to AI-enabled forms of automation, with some larger employers seeking to further automate tens of thousands of additional jobs through AI-powered robotics. The potential for mass unemployment in low-, middle- and high-income countries alike - along with the foreclosure of opportunity for many forms of labor previously seen as pathways to prosperity - threatens economic and political stability across the world. Capacity building, retraining, and upskilling may not prove sufficient to address these imminent gaps at scale. The profit potential of AI and automation for the private sector - and its detrimental impact on workers throughout the world - must be seen as an urgent threat. Unregulated data center proliferation is another issue that, if not adequately addressed in a timely manner, may prove impossible to slow. The impacts on the environment, air quality, freshwater resources, and carbon emissions from fossil fuel-powered data centers (which is the majority as of 2026), threaten to undermine all progress made towards climate change mitigation and adaptation. Thus far, data center construction has done little to provide economic growth to impacted communities and further threatens to increase the costs of energy access for everyday people and institutions they rely on, like schools and hospitals, while fossil fuel corporations are projected to earn nearly US$3,000 per second in 2026. Finally, while military AI was not in scope for the Global Digital Compact, any mitigation of the risks of AI in society (ranging from existential to group & individual risks) must apply equally to automated weapon systems.

The digital/AI rights movement comprises numerous under-funded CSOs and NGOs around the world pushing for rights-respecting standards, regulations, and commitments from the private sector, governments, and regulatory bodies. To-date, progress has been uneven, as a new era of international competition – characterized as an AI "arms race" with both geopolitical and commercial aspects – has emerged, with profit and global domination at the center. The current geopolitical climate of fear and confrontation has persuaded governments and private companies alike that they must be able to control the technologies on which their prosperity depends, placing blind faith in deregulation or techno-solutionism. Rather, as civil society and digital rights experts have long been advocating, well-crafted digital regulation and international cooperation around its enforcement can guide innovation in a direction that serves the public interest. The lack of a central visible, influential, and trustworthy voice to champion human rights in the era of AI places our movements in a persistent defensive posture. Funding is scarce, and opportunities for civil society to engage in complex digital governance environments are limited. As a result, the voices of the Global Majority remain underrepresented in digital governance discussions, which can perpetuate the neglect of urgent digital rights issues that disproportionately affect marginalised groups, for whom the risks - as well as the opportunities - are often greater. The Internet Governance Forum is an example of a process that promotes digital governance engagement on an equal footing for all stakeholders; the AI Dialogue should ensure that outputs from its discussions inform the IGF, to help close the loop between discussion and decision through a permanent UN forum. Narrowing or closing digital civic space and digital attacks on journalists and activists have further chilled the advocacy environment. The weaponization of AI threatens to make this situation permanent.

The UN Global Dialogue on AI Governance can play a meaningful role in advancing international cooperation on AI governance by serving as a trusted, inclusive space where diverse actors (particularly those historically excluded from global decision-making) and help shape shared norms and accountability. First, the Dialogue can help rebalance power. Today, AI decisions are often driven by a small number of governments and companies with questionable incentives and outsized resources. A well-designed Dialogue can centre the leadership, knowledge, and priorities of low- and middle-income countries, civil society, and impacted communities - recognising them not as beneficiaries, but as co-creators of governance frameworks. Second, it can build convergence around rights-based principles. Grounded in international human rights laws and standards, the Dialogue can support alignment on safeguards such as transparency, accountability, data protection, and remedies for harm. This is particularly urgent as AI systems increasingly shape access to livelihoods, public services, and civic space. A common foundation can help prevent regulatory fragmentation while ensuring that cooperation does not dilute protections. Third, the Dialogue can catalyze practical cooperation. Beyond principles, it can foster collaboration on issues such as cross-border data governance, public-interest AI and capacity sharing. This includes resourcing regulatory capacity in underfunded contexts and supporting locally-led innovation ecosystems. Cooperation should be designed to address structural inequalities, including the concentration of data and compute. Finally, the Dialogue must strengthen accountability. By creating mechanisms for ongoing learning, peer exchange, and public scrutiny, it can help ensure that commitments translate into practice. Civil society participation will be critical in this role. From Oxfam's perspective, the Dialogue's value will depend on whether it meaningfully shifts who sets the agenda, whose knowledge counts, and who benefits. Its success should be measured not only by consensus, but by whether it advances equality, opportunity, dignity, and agency in practice.

The UN Global Dialogue on AI Governance can build up on and connect across a growing ecosystem of initiatives, while addressing the fragmentation and power imbalances that limit their collective impact. The Independent International Scientific Panel on AI is clearly the cornerstone of shared understanding and it is Oxfam's hope that it will build a common vocabulary around risk probability and severity, as well opportunity, across civil society, private sector and governments, upon which issues can be discussed, and agreements and consensus, inclusive of impacted communities, can be built. Multiple other multilateral and multistakeholder initiatives can align and support the Global Dialogue, including the UN Global Digital Compact (particularly the progress reports), the UNESCO work on AI ethics, the OECD AI principles, the Global Partnership on AI, the AI Safety Summits, as well as regional governance and legislative frameworks. AI can't be contained by borders - regulation in one location is meaningless for humanity at large unless core human rights are guaranteed by all. The Global Dialogue is the legitimate space for such guarantees to be agreed and enforced. Ultimately the Global Dialogue must strengthen human rights in a digital age. By grounding collaboration in international human rights standards and feminist, anti-racist principles, it can help ensure that AI governance advances dignity, agency, and collective wellbeing of humans, and helps humanity mitigate the kind of catastrophic individual, collective and global risks that the current incentives and unbounded AI "arms race" are bestowing on us.

Different stakeholders can contribute to the Global Dialogue in ways that recognize their distinct roles, knowledge, and responsibilities, while ensuring that those most affected by AI systems have meaningful influence. For example: - Member States can share policy experiences, align on rights-based standards, and commit resources for equitable implementation; - Private sector actors can provide technical insight, increase transparency around emerging systems and impacts, and commit to accountability measures that go beyond voluntary principles; - Civil society, unions, and community-based organizations can surface lived experience, identify harms, and co-create solutions grounded in local realities; - Academia and technical experts can contribute independent research, interdisciplinary perspectives, and critical evidence, most especially from the Global Majority countries; - Affected communities must be engaged as decision-shaping participants through equitably-resourced and sustained participation. Recommendations for format and structure: - Multistakeholder with participation design to correct power imbalances e.g., dedicated seats, funding, and agenda-setting roles for under-resourced and under-represented groups; - Layered structure: Combine global convenings with regional and thematic dialogues, allowing context-specific issues and solutions to emerge and inform global outcomes; - Discussion-to-Decision pathways: Connect open discussions dialogue with working groups who are tasked with producing concrete outcomes (e.g. policy templates, accountability tools, or cooperation frameworks); - Transparency and accountability: Publicly document inputs, decisions, and follow-up actions, with mechanisms for independent monitoring and civil society oversight; - Accessibility and care: Ensure multi-language access, hybrid participation, and formats that enable meaningful engagement, recognizing barriers related to geography, resources, and time. From Oxfam's perspective, the Dialogue's legitimacy will depend on whether it redistributes voice and power, strengthens accountability, and translates participation into tangible, equality-oriented outcomes.

AI governance discussions continue to reflect unequal power and access and many voices remain consistently underrepresented, including: - Communities in global majority countries, particularly those who will be most affected by automation, resource extraction & climate vulnerability; - Workers across AI value chains, particularly those whose labour underpins AI systems but is often invisible; - Grassroots and community- based organizations, including feminist, and Indigenous movements, who bring critical knowledge about rights, care, and collective wellbeing; - Youth and future generations, whose entire lives and livelihoods will be shaped by governance decisions of today; This underrepresentation reflects structural inequality in funding, access, language, and agenda-setting power. Enabling meaningful inclusion would imply: - Providing resources for participation, rather than just invitations: i.e. funding, interpretation, travel support, and compensation for time and expertise, especially for grassroots actors and workers' representatives; - Promoting inclusive agenda-setting by creating formal roles for underrepresented groups to co-design priorities, not just to respond to them; - Elevating different forms of knowledge through recognising lived experience, community evidence, and Indigenous knowledge alongside technical expertise; - Using multi-layer engagement models by combining global forums with regional and community- level dialogues that ground decision-making in real-world perspectives; - Ensuring accountability pathways by creating mechanisms that ensure contributions shape outcomes, while ensuring participants can track how their input is used; - Addressing structural barriers by supporting long-term capacity building, including access to digital infrastructure, research funding, and policy spaces. From Oxfam's perspective, inclusion must go beyond representation toward redistribution of influence. The legitimacy of AI governance will depend on whether those most affected by AI systems can exercise agency in shaping them and whether cooperation actively dismantles, rather than reproduces, existing inequality.

Innovative engagement formats can help ensure the Global Dialogue moves beyond performative inclusion toward meaningful participation, shared learning, and collective accountability. Any approach that aligns with these outcomes would have to have these characteristics: - a commitment to human rights – for todays humans and those of the indefinite future – is woven throughout each stage; - a space where the least-heard but most-affected voices are elevated; - a process that brings together diverse perspectives in considering real-world examples of future scenarios; - an agenda that looks beyond the limited business models (and incentives) of today's technology landscape; - a format that promotes maximum accessibility, accommodating diverse means and channels of engagement; - a mechanism of transparency and traceable accountability for all outcomes that promotes trust and continued engagement.

Given the expertise you have already built, it seems unlikely that we can suggest anything to you that is not already well-known. Instead Oxfam encourages the Global Dialogue on AI Governance to choose approaches that centre the perspectives of affected communities, address structural inequality (where it exists in the present, and where it is already expected to emerge), and support the emergence of principles that can be concretely linked to enforceable accountability mechanism. The Global Dialogue on AI Governance can connect many existing and emerging efforts, close gaps in participation and implementation, and ensure governance advances dignity, agency, and equality.