Kairoi Responds to UN’s Call for Input
24th February 2025, by Ismael Kherroubi Garcia
Elise Racine & The Bigger Picture / Better Images of AI / Web of Influence I / CC-BY 4.0
On 21st February 2025, Kairoi submitted input for the terms of reference of the Independent International Scientific Panel (“the Panel”) on AI and the Global Dialogue on AI Governance (“the Dialogue”). The process is being co-facilitated by the permanent representatives to the United Nations of Costa Rica and Spain.¹ Both the Panel and the Dialogue are welcome advancements in the global governance of AI. The full response can be found online.² This post highlights Kairoi’s proposed mandates of both the Panel and the Dialogue.
Proposed Mandate for the Panel
To promote AI as an area of research as best performed through rigorous open science practices.
This mandate’s implications are twofold. On the one hand, the proposed mandate calls for reclaiming AI as a scientific discipline. Indeed, “AI” was coined in the halls of Dartmouth College almost 70 years ago. The succeeding decades saw peaks and troughs in interest and funding for AI research, and success stories often resulted from academic institutions such as MIT (e.g.: ELIZA) and Carnegie Mellon University (e.g.: R1). Meanwhile, the more recent advancements in AI have been led by for-profit organisations, such as Google, Meta, Microsoft and OpenAI. As exciting and valuable some people may find the resulting products to be, their creators’ motivations needn’t promote knowledge nor the rigour underpinning scientific research. Rather, for-profit organisations’ are driven by their shareholders’ financial interests. With this, the International Scientific Panel on AI has the opportunity to identify the institutional structures that are best positioned to advance impactful AI research. The assumption here is that for-profit organisations cannot be leading the space alone.
On the other hand, the mandate calls for good research practices as informed by the open science movement. Our work at Kairoi has shown time and time again —with organisations such as Open Life Science Limited, Next Generation Internet, the Open Modelling Foundation and GO FAIR US— that open science values and practices are powerful enablers of better AI research. Importantly, open science calls for an integration of good data management practices, multidisciplinary collaboration, community-driven standards, and technical solutions.³
Proposed Mandate for the Dialogue
To platform the voices of the general public, civil society and not-for-profit organisations before the for-profit AI innovators, policy-makers, and the International Independent Scientific Panel on AI.
With this mandate, the Dialogue is a platform for historically underrepresented voices to be heard by three major parties. Firstly, serving as a conduit between for-profit AI innovators and the general public —often, their potential customers—, the Dialogue can enable what the Responsible Artificial Intelligence Network (RAIN) calls “targeted” innovations; that is, innovations that respond to people’s actual needs, rather than for-profit motivations.⁴ Relatedly, meaningfully engaging with the public can help innovators foster trust.⁵
Secondly, it is worth highlighting that the three most prominent international AI-related summits —held in the UK in 2023, South Korea in 2024, and France in 2025— all ignited a sense that more must be done to include diverse voices in the AI space.⁶ ⁷ ⁸ In this regard, the proposed mandate would mark a drastic shift in current AI practices for the better. Indeed, such a mandate would allow the Dialogue to bring regulators much closer to the people for whom they are regulating.
Thirdly, by bringing the Dialogue and the Panel together in conversation, there is opportunity for the Panel to develop the technical assessment mechanisms that respond to the Dialogue’s concerns, and have oversight of their implementation across academia and industry. Conversely, the Dialogue can find in the Panel a source of independently and rigorously generated evidence concerning AI. Rather than depending on courses about AI developed by for-profit organisations, members of the Dialogue can build a degree of AI Literacy with the Panel.
The UN’s Growing Responsible AI Ecosystem
The above mandates would fit into a growing responsible AI ecosystem within the UN. An obvious tie with the Global Digital Compact is already present in how the Panel and the Dialogue are being designed. Going a step further, our own work for the International WELL Building Institute showed that responsible AI can be implemented through practical organisational mechanisms when linked to the UN Global Compact and the Sustainable Development Goals.⁹ The Internet Governance Forum is another relevant stakeholder in this space, and our proposal of a Panel mandate revolving around open science makes the case for linking up with the Global Open Science Partnership.
At Kairoi, we are excited to see how the Panel and Dialogue emerge and evolve, and we are pleased to see the UN continuing to lead the global discussion in responsible AI.
Contact us
hello@kairoi.uk
References
¹ UN (2025) AI Panel and Dialogue, UN Global Digital Compact, online [accessed 24 February 2025]
² Kairoi (2025) RE: UN Independent International Scientific Panel on AI and the Global Dialogue on AI Governance, Github, online [accessed 24 February 2025]
³ Kherroubi Garcia, I., Erdmann, C., Gesing, S. et al. (2025) Ten Simple Rules for Good Model-sharing Practices, PLOS Computational Biology, DOI: 10.1371/journal.pcbi.1012702
⁴ Kherroubi Garcia, I. et al. (2024) Introducing the Responsible Artificial Intelligence Network, online [accessed 21 February 2025]
⁵ Kherroubi Garcia, I. (2024) The Four Pillars of Responsible AI, Kairoi, online [accessed 21 February 2025]
⁶ Connected by Data et al. (2023) AI Safety Summit Open Letter to the Prime Minister, online [accessed 21 February 2025]
⁷ Klovig Skelton, S. (2024) AI Seoul Summit Review, Computer Weekly, online [accessed 21 February 2025]
⁸ Birhane, A. (2025) Bending the arc of AI towards the public interest, AI Accountability Lab, online [accessed 21 February 2025]
⁹ Pimentel, J. & Kherroubi Garcia, I. (2024) Leveraging Machines for Meaning: IWBI’s AI Framework for Advancing Healthy Buildings, IWBI, online [accessed 24 February 2025]
² Kairoi (2025) RE: UN Independent International Scientific Panel on AI and the Global Dialogue on AI Governance, Github, online [accessed 24 February 2025]
³ Kherroubi Garcia, I., Erdmann, C., Gesing, S. et al. (2025) Ten Simple Rules for Good Model-sharing Practices, PLOS Computational Biology, DOI: 10.1371/journal.pcbi.1012702
⁴ Kherroubi Garcia, I. et al. (2024) Introducing the Responsible Artificial Intelligence Network, online [accessed 21 February 2025]
⁵ Kherroubi Garcia, I. (2024) The Four Pillars of Responsible AI, Kairoi, online [accessed 21 February 2025]
⁶ Connected by Data et al. (2023) AI Safety Summit Open Letter to the Prime Minister, online [accessed 21 February 2025]
⁷ Klovig Skelton, S. (2024) AI Seoul Summit Review, Computer Weekly, online [accessed 21 February 2025]
⁸ Birhane, A. (2025) Bending the arc of AI towards the public interest, AI Accountability Lab, online [accessed 21 February 2025]
⁹ Pimentel, J. & Kherroubi Garcia, I. (2024) Leveraging Machines for Meaning: IWBI’s AI Framework for Advancing Healthy Buildings, IWBI, online [accessed 24 February 2025]
