ARTICLE 19, partner organisations, and international relations, justice, global governance, and technology experts have written to United Nations Permanent Representatives from Costa Rica and Spain regarding the 2 April 2025 UN Stakeholder Consultation on the Zero Draft of the AI Scientific Panel and Global Dialogue on AI Governance. Together, they urge Member States to address significant gaps in the draft to ensure that the Panel and Dialogue are grounded in human rights, meaningfully inclusive, transparent in process, coordinated with other relevant efforts, and capable of responding to the most serious risks posed by AI technologies.
The written submission is below.
Her Excellency, Maritza Chan Valerde
Permanent Representative of Costa Rica to the United Nations
His Excellency, Héctor José Gómez Hernández
Permanent Representative of Spain to the United Nations
Re: Written Follow-Up from 2 April 2025 UN Stakeholder Consultation on the Zero Draft of the AI Scientific Panel and Global Dialogue on AI Governance
Your Excellencies,
We, the undersigned organizations and individuals who have actively participated in the United Nations (UN) Global Digital Compact (GDC) process, welcome the opportunity to provide input on the Zero Draft of the Terms of Reference and Modalities for the Establishment and Functioning of the Independent International Scientific Panel on Artificial Intelligence and the Global Dialogue on Artificial Intelligence Governance (Zero Draft). We further thank you for the opportunity to supplement our oral interventions from the 2 April Stakeholder Consultation in writing.
We welcome the Zero Draft’s commitment to publish outputs in all UN languages, produce an annual reporting requirement for the Panel, and a Panel nomination process that includes consideration for geographic and gender balance, as well as financial, professional and personal interest disclosures. However, while these are important elements, we remain concerned that overall the text is too vague and open to discretionary interpretation; it, therefore, insufficiently captures key elements and the precision required for a transparent and concrete terms of reference and modalities necessary for the establishment and functioning of the Panel and the Dialogue. The following letter therefore addresses both our (1) procedural and (2) substantive concerns.
Procedural Concerns
Importantly, we are deeply concerned by the government-centric approach proposed in the resolution with regards to the Panel composition, particularly paragraphs 3(b), 4 and 5. While independence is required, there is a need for concrete strong safeguards to prevent corporate capture and/or state dominance, and to ensure meaningful participation from affected communities and the countries of the Global Majority. This should include making the disclosure of financial, professional, and personal interests that may affect impartiality or independence, not only part of the nomination process, but also an ongoing requirement for both Committees.
The structure of the Panel comprising two bodies — the Advisory Committee, elected by the General Assembly, and the Expert Committee — provides insufficient clarity on how the
Advisory Committee will relate to or influence the work of the Expert Committee. Paragraph 10 simply refers to the Advisory Committee’s contribution to the Panel’s outputs, without further clarification on the nature, timing and weight of this contribution. This lack of clarity risks creating an opaque and potentially politicised process, and could allow the Advisory Committee to mediate or neutralise the substance of the Panel’s outputs. Furthermore, it is truly a disincentive to have Experts participate on the Panel on a pro bono basis unless the Panel is truly independent. As we previously underscored during the 18 February consultation, the Panel must be ‘fully independent and free from political influence’. We, therefore, fear that there is no meaningful opportunity for multistakeholder engagement unless the Panel is fully independent and free from political influence. We maintain that this government-centric approach is a fundamental shift away from – and would further undermine – the existing inclusive and open multi stakeholder approach which has guided the work of the UN in digital governance for the past 20 years. We refer to the NetMundial+10 Principles on how to effectively and meaningfully operationalise multistakeholder participation.
The current mandate of the Dialogue is broad and risks overlapping with existing fora and processes unless its scope is better defined. We, therefore, wish to underscore that the Dialogue should complement existing AI policy dialogues and feed them into the UN Internet Governance Forum (IGF) as the main forum, as affirmed by the GDC. The purpose of the Dialogue is to facilitate discussion between those affected by AI and those developing AI to be in direct contact. We further echo concerns regarding the language in paragraph 12, particularly describing the Dialogue as both ‘multistakeholder’ yet also ‘intergovernmental.’ The Dialogue should not be an opportunity for member states to further regulate AI through intergovernmental negotiated outcomes as this would duplicate existing efforts on AI regulation within the General Assembly, ECOSOC, and the Security Council.
In addition, we cannot overlook the significance of paragraph 17, which calls to hold the first Dialogue in New York this September. This is deeply concerning in light of the current geopolitical context and specific limitations and risks faced by civil society from the countries of the Global Majority. We therefore call for a reconsideration of the location for any of the Dialogues to more inclusive and safe venues such as Geneva.
Along with others, we would welcome further clarity on the purpose and use of the Panel and Dialogue outputs. It remains unclear how these will be used, who they will inform, and how they will complement or reinforce existing governance initiatives. AI governance requires a holistic and global approach, turning a patchwork of initiatives into a coherent approach in compliance with international law, human rights and the UN Sustainable Development Goals. Lack of coordination and complementation could mean further fragmentation, impacting both coherence and participation. For both the Panel and the Dialogue there is an absence of a structured role for civil society – effectively the users that will be most impacted by AI technologies – nor any guarantee of meaningful participation beyond ambiguous references to ‘relevant stakeholders’ (paragraphs 4(a) for the Panel and 15 for the Dialogue). Without mechanisms for representation and resourcing, this risks reinforcing existing power imbalances and undermines the multistakeholder approach.
We further note the absence of any reference to coordination with existing UN human rights accountability mechanisms such as the UN Universal Periodic Review process, Special Procedures, or Treaty Bodies. We are deeply concerned that support for the Secretariat is limited to only one UN Secretariat backed UN entity, ODET, and two UN agencies – the ITU and UNESCO – without a clearly defined role for the Office of the High Commissioner for Human Rights (OHCHR).The GDC is anchored in human rights considerations and, therefore, both the AI dialogue and panel should ensure that human rights become the underlying foundation of their work.
Substantive Concerns
We insist that the mandates for both the Panel and the Dialogue be anchored in international law and the broad international human rights framework, including international human rights law, international humanitarian law and international refugee law. Human rights are mentioned just once at paragraph 14 in the Zero Draft related to the Dialogue. As a foundational principle, all human rights must be protected throughout the full lifecycle of all AI technologies. Concerningly, there is no reference to existing human rights frameworks such as the UN Guiding Principles in Business and Human Rights, or to the role of UN bodies like the OHCHR or the Human Rights Council. This is a major omission, particularly given the GDC’s mandate as well as the clear risks posed by many AI systems to the enjoyment of human rights. For example, the Zero Draft is silent on the Dialogue’s potential activities relating to high-risk AI uses such as predictive policing or emotion recognition, despite clear international concern and growing calls for prohibitions of AI uses that are incompatible with international human rights standards. In turn, while the Panel’s evidence-based role and multidisciplinary expertise is affirmed, there are no details on how human rights expertise will be integrated to the Panel and Dialogue.
We believe that for the Panel and Dialogue to be effective
they will need to balance scientific integrity with political legitimacy, including the nature of outputs. Protecting scientific integrity of the outputs of a scientific panel is paramount; however, appropriate political buy-in will be required to make this a success.This political buy-in can only be achieved through the human rights framework that will be able to guide any work the UN seeks to undertake with regards to AI.
In conclusion, we urge Member States to address these gaps to ensure that the Panel and Dialogue are grounded in human rights, meaningfully inclusive, transparent in process, coordinated with other relevant efforts, and capable of responding to the most serious risks posed by AI technologies. Overall, we are alarmed that the Zero Draft, as worded, would set a dangerous precedent which could further legitimize repressive digital practices, hinder innovation of inclusive economic growth, diminish the decentralized nature of the internet, and undermine human rights.
SIGNATORIES (alphabetically ordered)
Organizations
Access Now
ARTICLE 19
Association for Progressive Communications
Derechos Digitales
European Center for Not-For-Profit Law Stichting
Global Partners Digital
International Center for Not-for-Profit Law
Individuals
Dr Konstantinos Komaitis, Resident Senior Fellow, Global Governance and Technology Lead, Democracy and Tech Initiative, Atlantic Council
Dr Katharine Millar, Associate Professor, Department of International Relations, London School of Economics
Dr Sebastian Smart, Senior Fellow, Centre for Access to Justice & Inclusion, Anglia Ruskin University