Digital Rights Foundation

For the first Global Dialogue on AI Governance to be considered a success, it must move beyond aspirational declarations and produce concrete, rights-based outcomes that reflect the realities of communities most harmed by ungoverned AI systems, particularly those in the Global South. A meaningful outcome would be the establishment of an inclusive, ongoing governance mechanism that ensures equitable participation from Global South countries, civil society, and marginalised communities in future AI policy processes. A one-time dialogue is insufficient; success requires building a sustained architecture for representation and accountability. The Dialogue should produce clear, actionable recommendations on the most urgent AI harms, including AI-amplified technology-facilitated gender-based violence, synthetic media and disinformation, and the use of AI in conflict contexts. For organisations like DRF working at the intersection of AI and human rights in South Asia, success means seeing these harms explicitly named, documented, and addressed within governance frameworks, not treated as secondary concerns to economic or technical priorities. Outcomes should also include commitments to transparency and human oversight in AI systems, particularly by private-sector actors. Platform accountability mechanisms that are accessible to users in low-resource settings, available in local languages, and responsive to context-specific harms would represent a significant step forward. Critically, success requires moving away from governance models that replicate Global North priorities and toward frameworks that are contextually relevant, intersectional, and grounded in international human rights standards. This means recognising the digital gender divide, data colonialism, and language exclusion as governance issues, not peripheral concerns. Ultimately, the Dialogue succeeds if it produces outcomes that those most affected by AI harms, women, minorities, and vulnerable communities across the Global South can point to as genuinely improving their safety, rights, and participation in digital spaces.

• 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

In South Asia, the rapid adoption of AI technologies, especially generative AI, has significantly reshaped information ecosystems. AI-driven disinformation, including synthetic media, is increasingly being used not only by individuals but also by state and non-state actors, particularly in conflict contexts. This has serious implications for information integrity, public trust, and human rights, often exacerbating existing inequalities and disproportionately impacting women, minorities, and vulnerable communities. Digital Rights Foundation (DRF) prioritises these thematic areas based on its extensive work at the intersection of AI, human rights, and digital governance in the Global South. DRF has actively contributed to research and advocacy on the social, technical, and ethical implications of AI, particularly in relation to technology-facilitated gender-based violence (TFGBV), disinformation, and platform accountability. DRF's work highlights the urgent need for safe, secure, and trustworthy AI systems, alongside stronger transparency, accountability and human oversight mechanisms. Current AI governance frameworks are largely shaped by Global North priorities and fail to adequately address the lived realities and risks faced in the Global South, including weak regulatory environments, limited platform accountability, and heightened risks of misuse in politically sensitive contexts. DRF also emphasises the protection and promotion of human rights as central to AI governance, ensuring that emerging technologies do not undermine freedoms of expression, privacy and access to information. At the same time, understanding the social, economic, and cultural implications of AI is critical to developing inclusive and contextually relevant governance models. DRF's selection reflects the need for urgent, inclusive, and rights-based global dialogue that incorporates Global South perspectives, addresses real-world harms and ensures that AI systems are developed and deployed in ways that are equitable, accountable, and aligned with international human rights standards.

Yes, several cross-cutting and emerging issues remain insufficiently captured within the listed themes, particularly from a Global South perspective. First, the use of AI in conflict and geopolitical contexts is an urgent and under-addressed issue. Across South Asia and other regions, AI, especially generative AI, is increasingly being used in information warfare, including cross-border disinformation, synthetic media manipulation and narrative control. These uses blur the lines between civilian and military applications of AI and have direct consequences for human rights, peace and democratic processes. Second, technology-facilitated gender-based violence (TFGBV), as amplified by AI systems, requires more explicit recognition. AI tools are being used to generate non-consensual intimate imagery, automate harassment, and scale gendered disinformation campaigns. These harms disproportionately target women, journalists and marginalized communities, leading to silencing and exclusion from public discourse. Moreover, there is a growing concern around data colonialism and extractive data practices, where data from the Global South is used to train AI systems without meaningful consent, benefit-sharing or contextual safeguards. This raises questions of equity, ownership, and whose knowledge systems are represented in AI development. Additionally, language and cultural exclusion remain a critical gap. Many AI systems lack adequate support for local and low-resource languages, leading to biased moderation, misclassification of harmful content and exclusion of large populations from digital participation. The lack of accessible accountability mechanisms for users, particularly in the Global South, limits the ability to challenge harms caused by AI systems and platform decisions.

Governance gaps in AI are significantly shaping the information ecosystem, human rights landscape and media environment in South Asia and Global South. One of the most pressing challenges is the rapid proliferation of AI-driven disinformation, including synthetic media, which is being used in political discourse, cross-border tensions and conflict settings. The absence of clear regulatory frameworks and limited platform accountability has made it difficult to effectively respond to these harms, undermining information integrity and public trust. These risks are further heightened by weak enforcement of existing laws and the misuse of legal frameworks, such as cybercrime regulations, which can disproportionately target journalists, activists, and dissenting voices. This creates a dual challenge with harmful content spreading largely unchecked and legitimate expression is often curtailed, contributing to a shrinking civic space. Women and marginalized communities are particularly affected, as AI tools amplify TFGBV, including deepfakes, harassment and coordinated disinformation campaigns. Another key gap is the lack of transparency and human oversight in AI systems used by both state and private actors. Limited disclosure around how algorithms function, how content is moderated or how decisions are made restricts accountability and user recourse. Additionally, Global North-centric AI governance models fail to address local realities, including language diversity, socio-political contexts and infrastructural constraints, leading to ineffective or exclusionary outcomes in the region. However, these challenges also present opportunities. There is growing awareness among civil society, policymakers and media actors about the need for rights-based AI governance. This creates a space for multi-stakeholder collaboration to develop context-sensitive frameworks, strengthen digital literacy and build capacity among journalists and communities. With targeted investment and inclusive policy development, AI can also support innovation in media, improve access to information and enhance civic engagement in the region.

Since the AI Dialogue allows every country a seat at the table, one can hope for an equitable dialogue among the Global Majority countries, as well as dialogue between these countries and Global Minority countries. International cooperation is built on international dialogue, of the kind that is detailed and takes into account the highly varying risks and threats posed, plus the opportunities offered by AI technologies in various regions of the world. For example, in South Asia, the digital gender divide is baked into conversations on AI and AI governance, whereas regions like the US are running into issues such as job displacement, while one of Europe's biggest concerns is regulatory compliance. It is through international cooperation on a large scale that these problems and their solutions can be sufficiently fleshed out by country representatives and experts, and the AI Dialogue presents a promising opportunity. Since the Dialogue builds on the Independent International Scientific Panel on AI reports, it ensures that these findings are shared with policymakers and tech investors worldwide, ostensibly helping them make more informed decisions about AI policy and investments. Not only this, but since AI decision-making, especially pertaining to policy, often happens in silos, one potential outcome of the Dialogue is a meeting of minds between policymakers, tech experts, academics, researchers, civil society, and other stakeholders, who each bring their specific expertise to the table.

The AI Dialogue can build on existing global, regional, and multistakeholder initiatives shaping AI governance. Key UN efforts include the UN AI Advisory Body's 2024 report, the Global Digital Compact, and the UN AI Governance White Paper, alongside UNESCO's Ethics of AI Recommendation. Legally binding frameworks like the Council of Europe AI Convention and policy standards such as the OECD AI Principles offer regulatory foundations. Multistakeholder platforms like the Global Partnership on AI, as well as regional strategies (EU, ASEAN, African Union), and initiatives like the G7 Hiroshima Process further support coordination, capacity building, and alignment across diverse governance approaches. The Global Network Initiative (GNI) is a multistakeholder collective which brings together civil society, tech companies, human rights advocates, investors, and academics to work on issues pertaining to privacy rights, tech governance, and freedom of speech and expression. The Digital Accountability Collective South Asia (DACSA) is a South Asia-based collective working on the risks associated with existing and emerging technologies, tech accountability, and tech governance. Bringing inclusive and representative collectives such as this to the table will ensure that the conversations being had are substantive and discuss tangible issues. The AI Dialogue can add value to these initiatives and collectives by expanding the scale at which collaborations are happening. However, this can only be achieved if the conversations indeed carry weight, especially in pragmatic terms, and the Dialogue is inclusive of all voices from all regions.

Effective AI Dialogue requires meaningful, inclusive and sustained engagement from a wide range of stakeholders, with particular emphasis on representation from the Global South and communities most impacted by AI systems. Governments should contribute by committing to transparent, rights-based policymaking, sharing regulatory experiences, and engaging in cross-border cooperation, especially on issues such as disinformation and the use of AI in conflict contexts. Civil society organisations play a critical role in bringing forward lived experiences, documenting harms and advocating for human rights-centred approaches, particularly for marginalised communities. The private sector, including technology companies, must engage by improving transparency, sharing relevant data, strengthening accountability mechanisms and ensuring responsible development and deployment of AI systems. Academia and technical experts can provide interdisciplinary research and risk assessments to support evidence-based policymaking, while media and journalists are essential in shaping public discourse, raising awareness, and highlighting emerging risks. In terms of format and structure, the AI Dialogue should adopt a multi-stakeholder and regionally inclusive approach, ensuring equitable participation from Global South actors, grassroots organisations, and underrepresented voices. Regional consultations should feed into global discussions to ensure context-specific insights are reflected. The Dialogue should be organised through thematic working groups focusing on areas such as human rights, disinformation, conflict and accountability, with clear and actionable outputs from each group. A hybrid format combining in-person and virtual participation is essential to ensure accessibility, including language inclusivity and support for participants from low-resource settings. Moreover, the Dialogue should be action-oriented, producing tangible outcomes such as policy recommendations, best practice frameworks and accountability benchmarks.

Conversations on AI governance and work needed to tackle emerging AI-threats and harms are largely dominated by the West - including their governments, Big Tech firms and other stakeholder institutions. As a result, multiple voices from the Global South and marginalised communities go unheard, misunderstood, and unaddressed. Regions such as Africa, South Asia and East Asia are increasingly becoming the hub of AI-driven data centres that use up vital resources from these scarce areas and yet are still excluded from conversations around global AI governance. Marginalised groups such as women, children, the disabled, and individuals from rural communities – mostly found in the Global South – and their unique struggles with AI either in personal or professional lives are similarly not reflected in policy discussions or frameworks. Another vastly overlooked demographic is individuals from different industry sectors who, indirectly, are being affected by AI. Additionally, interdisciplinary perspectives beyond the technology sector, such as academics and civil society organisations that work on human rights, are underrepresented in these spaces, giving discussions a one-dimensional aspect. Advancements in AI technologies adversely affect the most vulnerable in societies around the world, and those are the exact people who lack the space to raise awareness for the harms they face in digital spaces governed by AI. Efforts made to fund and support regional participation from underrepresented areas should be prioritised. This involves not only including civil society organisations but also community representatives from marginalised sectors to ensure that discussions are grounded in real-world contexts.

Discussions should be rooted in making conversations accessible and easy for every participant involved to ensure they are heard, understood and given the opportunity to share their perspectives. Firstly, panel discussions should prioritize a diverse set of experts from every region. Policy makers from Global South countries should also be invited in order to fully grasp the importance of creating holistic AI policy frameworks. To engage with disabled communities, provisions should be made for assisted technology and sign language interpreters to allow them to fully participate in the dialogue. Similarly, translators for regional languages should be available at the meetings to allow participants to engage in easy, meaningful and efficient ways. Lastly, hybrid collaborations should be allowed to ensure participants from underprivileged backgrounds and with weak passports can be guaranteed attendance if visas are not processed on time.

There are multiple international guidelines around artificial intelligence (AI) that list standards of use for governments around the world, focusing on . These include OECD AI principles, EU AI Act and the United Nation's guidelines on AI use. While comprehensive in their outlook, international charters around the use of AI within governments are not enough to tackle potential AI harms. Civil society organizations from across the globe including Access Now, the Business and Human Rights Resource Center, and Article 19 have also outlined principles for AI use within governments, business and institutions, focusing solely on minimizing the human rights impact from the use of such tools as well as prioritizing transparency and accountability when it comes to deploying AI tools across public and private sectors. Additionally, civil society offers an interdisciplinary approach to charting our AI guidelines that not only look at what regulations need to be in place for AI to function within society safely but also how the environment, education systems, or processes such as immigration can be adversely affected by the proliferation of AI. Whilst combining civil society and government guidelines gives a broad and far reaching outlook to emerging AI rules and regulations, it's important to also include perspectives from the Global South to ensure that challenges unique to the region, and those disproportionately affected, can be addressed.