Center for AI and Digital Policy (CAIDP)

An indicator of success for the Global Dialogue on AI Governance is to produce concrete, operational, and effective implementation initiatives that protect and promote human rights. This includes: (1) The establishment of a UN Special Rapporteur on AI and Human Rights to horizontally cover all human rights elements, provide leadership and expertise to define international red lines for AI systems that violate human rights and pose unacceptable risks, and act as an independent reporting mechanism. (2) Bridge AI safety and fairness agendas to truly advance reliable and trustworthy AI systems. CAIDP cautions against narrowly framing AI safety only as a technical or existential risk. AI systems are socio-technical systems with significant impact on people's lives, liberties, and access to resources and opportunites. To build trust, we need both technical guardrails and protection of human rights, particularly in high-impact public sector and consumer-facing systems. Innovation must advance our rights and well-being. (3) Promote explicit red lines and termination obligation on pseudoscientific and fundamentally discriminatory AI systems (such as emotion analysis, biometric categorization, biometric mass surveillance, predictive policing, and those that undermine children's safety. These mentioned systems are already banned in several jurisdictions, and there is a growing global convergence on such redlines. This should also serve as a baseline for the Independent International Scientific Panel on AI's work.

• 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

CAIDP's mission is to promote a better society, more fair, more just-a world where technology promotes broad social inclusion based on fundamental rights, democratic institutions, and the rule of law. The above mentioned priorities are also priorities evaluated in the CAIDP AI Index, which assesses national AI policy and practices of 90 countries against their commitment and implementation of human rights and democratic principles. The rapid advancement of modern AI has created a profound governance challenge for political leaders, policymakers, civil society organizations, and citizens around the world. AI's novel capabilities and potential benefits have encouraged governments and business leaders to accelerate its development and deployment. Yet at the same time, the rapid advancement and widespread adoption of AI without guardrails have introduced serious risks to human rights, democratic processes, and the rule of law. Significant gaps exist between the principles governments commit to and how they implement (or lack implementation) those commitments. Some of the most critical elements assessed in the CAIDP AI Index are whether a government engages the public in AI policymaking, requires mechanisms for algorithmic transparency, and whether it meets its obligations for fairness, independent oversight, accountability, and other protections for its citizens. Drawing from the findings of the CAIDP AI Index, we provide expert guidance to governments and multilateral organizations on AI governance that is grounded in human rights, democratic values, and the rule of law. The CAIDP AI Index provides a key global reference resource for policymakers, academics, and civil society, and is usedin AI literacy initiatives to build capacity among civil servants and train the next generation of AI governance leaders. CAIDP's alumni serve in civil society organizations, governments, international institutions and businesses across 130 countries where they help shape the future of global AI governance.

An emerging issue not explicitly captured by the listed themes above are the human rights implications of Agentic AI technologies, and the governance safeguards required to ensure fundamental rights are not violated. AI systems with agentic capabilities can take actions on behalf of human actors, and can be misused at scale and speed. Agentic AI can tailor personalized disinformation and target communities. Agentic AI can aggregate information and provide granular information to undermine privacy, righto to expression and association.CAIDP has provided policy expertise to the G7 Digital and Technology working group, OECD and GPAI on agentic AI.

Public initiatives such as AI Incident database, AI Litigation database, or AI, Agorithmic and Automation Incidents Repository provide many examples across jurisdictions. We are not lacking examples or evidence of how governance gaps impact us. What is missing is implementation of oversight and accountability. The CAIDP AI Index assessment of 90 countries reveal a number of governance and implementation gaps, including the public participation in policymaking, absence of clear red lines on unscientific systems, a lack of termination obligation on unsafe AI, insufficient algorithmic transparency, and independent oversight of AI. Global Dialogue can take the Index's findings as a starting point.

The Global Dialogue should move discussions toward demonstrable impact grounded in human rights, democratic values, and the rule of law. A key step would be to help catalyze the establishment of a UN Special Rapporteur on AI and Human Rights. Such a mandate would give the United Nations the agility needed to respond to the fast-moving pace of AI development, strengthen its authority to engage credibly with governments, industry, and civil society, and provide the specialized competence needed to assess both the technical and human rights dimensions of AI. The Special Rapporteur would provide independent reporting to strengthen the work of the Global Dialogue, as well as the Independent International Scientific Panel on AI, while helping establish international red lines for AI systems that violate human rights and pose unacceptable risks. Through this mechanism, the Global Dialogue can translate international cooperation into concrete outcomes on the most urgent global priorities, such as prohibiting AI-enabled mass surveillance, safeguarding the right to explanation in automated decision-making, and protecting vulnerable populations from AI systems that entrench systemic bias and inequality.

The Global Dialogue should leverage the CAIDP AI Index as an authoritative global resource for assessing AI governance, tracking implementation of international commitments, identifying gaps, and prioritizing areas for urgent international cooperation. The CAIDP AI Index can help guide Member States toward policies aligned with human rights, democratic values, and the rule of law. It offers the United Nations with a clear, evidence-based comparative analysis of national AI policies, highlighting progress, backsliding, and emerging trends. The Global Dialogue should promote the critical role UNESCO's Recommendation on Ethics of AI and its Readiness Assessment Methodology (RAM) in centering human rights in AI governance. Dozens of countries have now completed or are in the process of RAM to identify their specific gaps and establishing a roadmap. The Global Dialogue should also connect with CAIDP's AI Policy Clinics to advance AI literacy and capacity building among Member States. CAIDP's AI Policy Clinics are designed to equip policymakers, leaders, civil servants, civil society organizations, academia, and the private sector with the necessary AI literacy and expertise to govern AI effectively.

The Global Dialogue should enable meaningful and sustained participation from governments, civil society, academia, and industry, with clearly defined roles in agenda-setting, policy development, and evaluation. CAIDP has emphasized the need for transparent, democratic AI governance processes, including meaningful civil society participation and reporting on AI impacts. Stakeholders should contribute directly to governance outcomes. Participation should not be limited to consultation. The Global Dialogue should therefore establish thematic working groups with inclusive and balanced geographic and gender representation, particularly participants from the Global Majority countries. Such working groups should produce concrete outputs, including policy recommendations, governance benchmarks, and implementation guidance. The Global Dialogue should also promote UNESCO's efforts in capacity building and knowledge sharing, particularly for countries with limited capacities and resources. Without such support, participation risks becoming symbolic rather than substantive.

Global AI governance discussions continue to underrepresent stakeholders from the Global Majority countries. CAIDP has highlighted the importance of equitable participation and decision-making in global AI governance, particularly for the Global Majority. CAIDP has trained more than 2,000 professionals in AI policy and governance. Global discussions can be inclusive when there is real intention. Efforts to ensure meaningful participation must be supported by capacity building, a priority consistently highlighted by Member States from the Global Majority. CAIDP has also emphasized that AI governance must be inclusive, meaningful, and representative of affected individuals. To support this, targeted inclusion and accessibility measures are needed.

The Global Dialogue should enable greater participation from civil society organizations representing a broad range of stakeholders and perspectives. While Member State authorities are essential for operationalizing outcomes from the discussions, civil society organizations provide critical on-the-ground insights into the human rights impacts of AI systems and are important actors in reinforcing accountability for commitments made in the Global Dialogue. CAIDP has emphasized that effective AI governance requires engagement processes that are transparent, democratic, and outcome-oriented ensuring key stakeholders are meaningfully and directly informing decisions. The Global Dialogues should adopt structured and iterative formats that combine deliberation, evidence generation, and accountability. This includes three core formats: First, multistakeholder deliberative panels that bring together policymakers, technical experts, civil society, and affected communities to assess specific AI risks and develop actionable policy options under defined mandates and time-bound cycles. Second, thematic working groups with clear mandates and structured written submission processes, supported by iterative review cycles and publicly documented outputs. Third, transparent consultation mechanisms, including publication of draft recommendations, structured public comment periods, and clear documentation of how inputs are reflected in final outcomes. Together, these formats would help ensure that participation in the Global Dialogue is not only broad, but structured, iterative, and demonstrably linked to outcomes in AI governance.

In providing expert guidance to governments and multilateral organizations on AI governance, CAIDP has consistently advocated for approaches that uphold and promote human rights, democratic values, and the rule of law. To support our policy recommendations, we draw on the following key frameworks and initiatives: ● The CAIDP AI Index is the most comprehensive global assessment of national AI policies and practices evaluated against human rights and democratic values. ● The Universal Guidelines for AI (UGAI), established in 2018 by diverse stakeholders, is the first rights-based AI guideline to promote transparency and accountability, with obligations for organizations toretain control over the systems they create. ● The UNESCO Recommendation on the Ethics of AI places protection of human rights and dignity at its core, and is grounded in principles such as transparency and fairness. ● The Council of Europe Framework Convention on AI, Human Rights, Democracy and the Rule of Law is the world's first legally bringing AI treaty that ensure activities within the AI lifecycle are consistent with human rights, democracy, and the rule of law. ● The EU AI Act offers a risk-based regulatory framework that imposes corresponding obligations to protect fundamental rights, ensure transparency, and accountability. ● The Hiroshima AI Reporting Framework serves as a voluntary mechanism encouraging transparency and accountability among developers of advanced AI systems. ● Human Rights-based public procurement guidelines that ensure safe AI adoption in the public sector and serve as a powerful policy lever for shaping AI development that embeds safeguards for human rights, transparency, and accountability.