Global Initiative for Inclusive Information and Communication Technologies (G3ict)

A successful Global Dialogue would produce clear, actionable, and rights-based foundations for international AI governance. First, it should establish a shared global understanding that accessibility, non-discrimination, and human rights—particularly the rights of persons with disabilities—are core governance principles, not optional safeguards. This includes explicit recognition of accessibility standards, inclusive design, and the need for representative datasets and human oversight. Second, the Dialogue should generate interoperable governance pathways that help countries align emerging AI regulations, standards, and risk-management approaches. This is especially important for low- and middle-income countries that face capacity constraints and risk being left behind by fragmented global frameworks. Third, the Dialogue should catalyze capacity-building commitments, including technical assistance, knowledge transfer, and support for accessible digital public infrastructure. This should include mechanisms for strengthening national institutions, procurement systems, and regulatory bodies to ensure AI systems are safe, trustworthy, and inclusive. Finally, success requires meaningful participation of underrepresented groups—especially persons with disabilities and their representative organizations (OPDs)—in both the Dialogue and the governance structures it informs. Embedding inclusive participation from the outset will set a precedent for future global AI governance processes.

• Safe, secure and trustworthy AI
• AI capacity-building
• Interoperability of governance approaches
• Protection and promotion of human rights

G3ict prioritizes these four areas because they directly determine whether AI systems advance or undermine the rights of persons with disabilities and other marginalized groups. Safe, secure, and trustworthy AI is essential because inaccessible or biased AI systems can automate discrimination, restrict access to services, and create new barriers. Ensuring safety and trustworthiness requires embedding accessibility, inclusive datasets, and human oversight into system design and deployment. AI capacity-building is critical for governments, especially in the Global South, to implement accessibility standards, conduct impact assessments, and procure inclusive AI systems. Without capacity, even strong policies cannot translate into equitable outcomes. Interoperability of governance approaches matters because fragmented global frameworks create uncertainty and widen inequality. Harmonized standards-particularly around accessibility, data governance, and risk management-enable countries to adopt inclusive AI without duplicative or conflicting requirements. Protection and promotion of human rights is foundational. AI governance must align with the CRPD and other human rights instruments to ensure equal access, non-discrimination, and participation. This includes ensuring AI systems used in public services, employment, education, and health do not create or reinforce exclusion. Together, these four areas form the minimum enabling environment for accessible, inclusive, and rights-based AI ecosystems.

Several cross-cutting issues require explicit attention. First, digital accessibility and inclusive design remain insufficiently recognized as core governance requirements. Accessibility is often treated as a technical afterthought rather than a structural obligation grounded in human rights law. Without explicit inclusion, AI systems will continue to exclude persons with disabilities. Second, data poverty and representational bias-particularly affecting persons with disabilities, older persons, linguistic minorities, and low-resource communities-pose systemic risks. AI governance must address inclusive data collection, ethical data stewardship, and safeguards against exclusionary datasets. Third, AI in digital public infrastructure (DPI) is an emerging area with profound implications. As governments integrate AI into identity systems, payments, health, and social protection, accessibility and rights protections must be embedded from the outset. Fourth, procurement and market incentives are powerful but underutilized levers. Public procurement can drive adoption of accessible and trustworthy AI, but only if standards and requirements are harmonized and enforceable. Finally, global inequities in compute, talent, and infrastructure risk deepening divides. Governance must address equitable access to AI resources and ensure that low- and middle-income countries can meaningfully participate in shaping global norms.

Governance gaps in accessibility, human rights protections, and interoperability directly affect digital inclusion efforts worldwide. In many countries, AI systems used in public services—such as identity verification, hiring, education, and health—are deployed without accessibility requirements or disability impact assessments. This results in exclusion from essential services and reinforces structural inequalities. Lack of capacity-building means governments often cannot evaluate AI systems for accessibility or bias, leaving persons with disabilities disproportionately affected by errors, misclassification, or inaccessible interfaces. Fragmented governance approaches also create uncertainty for industry, slowing adoption of accessibility standards and limiting the availability of inclusive AI tools globally. These gaps hinder progress toward CRPD implementation and undermine digital public infrastructure initiatives that could otherwise expand access and participation.

The AI Dialogue can serve as a neutral, multilateral platform that bridges fragmented governance efforts and accelerates global alignment. It can articulate shared principles grounded in human rights, including accessibility and non-discrimination, and promote interoperability across national and regional frameworks. The Dialogue can also strengthen capacity-building ecosystems, connecting countries with technical expertise, standards bodies, and international organizations. This is especially important for ensuring that low- and middle-income countries can adopt inclusive AI governance models and participate in global standard-setting. Additionally, the Dialogue can institutionalize inclusive participation, ensuring that persons with disabilities, OPDs, civil society, academia, and the private sector contribute to shaping global norms. By embedding inclusive governance from the outset, the Dialogue can set a precedent for future international cooperation. Finally, it can act as a coordination hub, linking existing initiatives—such as the Global Digital Compact, OECD AI Principles, UNESCO AI Ethics Recommendation, and ITU accessibility standards—to avoid duplication and promote coherence.

The Dialogue should connect with and build upon several established frameworks and partnerships: · UN CRPD, particularly Articles 4, 9, and 21, which provide binding obligations on accessibility and non-discrimination. · Global Digital Compact, which emphasizes inclusive digital public infrastructure and rights-based digital governance. · OECD AI Principles and UNESCO AI Ethics Recommendation, which offer widely adopted normative foundations. · ITU accessibility standards and W3C/WAI guidelines, which provide technical specifications for accessible digital systems. · Digital Public Goods Alliance and ID4D, which support inclusive digital public infrastructure. · DARE Index and G3ict's global benchmarking initiatives, which provide data-driven insights on accessibility readiness. The added value of the AI Dialogue lies in its ability to synthesize these efforts, promote interoperability, and elevate accessibility and human rights as cross-cutting priorities across all AI governance domains.

Governments should contribute by sharing regulatory approaches, national strategies, and lessons learned from deploying AI in public services. They should also commit to integrating accessibility and human rights safeguards into national AI frameworks. Civil society and OPDs should provide lived-experience expertise, monitor impacts, and co-design governance mechanisms. Their participation must be resourced and institutionalized, not symbolic. Industry should contribute technical insights, adopt accessible-by-design practices, and align products with global standards. Academia should support evidence generation, impact assessments, and interdisciplinary research. The Dialogue's structure should include: · Thematic working groups with balanced stakeholder representation. · Regional consultations to ensure Global South leadership. · Accessible hybrid participation formats to enable global engagement. · A standing advisory group of persons with disabilities and OPDs to ensure continuous inclusion. · Mechanisms for public reporting and accountability.

Persons with disabilities and their representative organizations remain significantly underrepresented in global AI governance. Their inclusion is essential to ensure AI systems do not perpetuate exclusion or discrimination. Other underrepresented groups include older persons, Indigenous communities, linguistic minorities, people in low-resource settings, and women and girls with disabilities. Many lack access to policymaking spaces, digital infrastructure, or the resources needed to participate meaningfully. Inclusion requires: · Dedicated funding for participation of OPDs and marginalized groups. · Accessible formats, including captioning, sign language interpretation, easy-to-read materials, and screen-reader-friendly documents. · Regional and community-based consultations to capture diverse perspectives. · Institutionalized representation, such as advisory bodies or co-chair roles for underrepresented groups. · Capacity-building to support meaningful engagement in technical and policy discussions.

Effective engagement requires formats that are participatory, accessible, and globally inclusive. These may include: · Accessible hybrid roundtables with real-time captioning, sign language interpretation, and multilingual support. · Scenario-based simulations that allow stakeholders to explore governance challenges and test policy options. · Inclusive design sprints bringing together policymakers, OPDs, technologists, and industry to co-create solutions. · Regional innovation labs focused on AI in public services, digital public infrastructure, and accessibility. · Open consultations on draft governance frameworks, using accessible digital platforms. · Youth and OPD-led dialogues to elevate perspectives often excluded from global processes. These formats can deepen participation, surface diverse expertise, and strengthen legitimacy.

Several initiatives offer promising models: · The EU AI Act's accessibility provisions, which require accessible-by-design systems and human oversight. · Canada's Algorithmic Impact Assessment, which integrates rights-based risk evaluation into public-sector AI deployment. · Singapore's Model AI Governance Framework, which provides practical guidance for industry adoption. · W3C/WAI accessibility standards and ITU-T accessibility guidelines, which offer mature technical foundations for inclusive AI. · Digital Public Goods Alliance projects that embed accessibility into digital public infrastructure. · G3ict's DARE Index and AI/DPI Accessibility Module, which provide global benchmarking and actionable pathways for improving national readiness. These examples demonstrate that effective AI governance requires a combination of rights-based principles, technical standards, institutional capacity, and inclusive participation.