Despite covering more than 190 member states and having the world’s leading states, politically and economically, at its helm, the UN has still not released any formal regulation or agreement on artificial intelligence yet.

By Nandini Parashar

Artificial Intelligence (AI), a technology that collects, classifies, and identifies patterns in data to perform tasks that are usually done by humans, could well be the most prominent buzzwords of the twenty-first century. With constant technological innovations such as ChatGPT, DeepSeek AI, etc., AI has found top priority among all important actors in world politics. A prime example is the recent AI Roadmap released by India and France that set the tone for harnessing AI’s potential through innovation and trade.

Across the world, countries and supranational organisations like the EU have also started gearing for AI regulations through laws, ethics documents and strategy plans. However, among all these developments, one corner from where a surprising silence emerges is the United Nations. Despite covering more than 190 member states and having the world’s leading states, politically and economically, at its helm, the UN has still not released any formal regulation or agreement on artificial intelligence yet.

It might not, pertinently, be accurate to say that the UN has not picked up on such ongoing technological developments at all – at the institutional level, the UN has created the Inter – Agency Working Group on AI and the Office for Digital and Emerging Technologies.

AI Governance and Adoption

However, the UN’s approach to AI can be classified under two headings – firstly, the adoption of AI by UN Systems for their internal functioning; and secondly, the adoption of AI by member –  states and other institutions. Out of the two, the UN has been actively pursuing the former and has recognised the need to adopt AI within its own structures and bodies to increase efficiency and enhance productivity.

One of the first few guiding documents for the same is a UN System-wide strategic approach and roadmap for supporting capacity development for AI. It highlights the key goals and themes for the UN Systems in their work towards AI governance and adoption, like, data, infrastructure, human and social capabilities, human rights and international law, etc. Further, at the Joint Session of the High-Level Committee on Programmes (HLCP) and the High-Level Committee on Management (HLCM) in 2023, a Task Force was also set up to develop system-wide guidelines and model policies on the ethical use and adoption of AI within UN Systems.

Through such and other publications like the Report on the Operational Use of AI in the UN System (2024), The UN Sec-Gen’s Roadmap for Digital Corporation (2018), UN 2.0 Quintet for Change; UN Sec. Gen’s Data Strategy for Action by Everyone, Everywhere with Insights, Impact, and Integrity: 2020 – 2022, the UN has identified several values and action plans that further the idea of Do No Harm, such as decentralised workings, creating ethical use guidelines for AI and data sharing, risk identification and assessment and so on. The issue of Ethical Impact Assessment tools, within UN Systems and beyond, has subsequently been taken up by UNESCO.

Summit for the Future

On the other hand, despite consistent AI governance around the world, the UN’s work on regulating AI adoption among member-states remains anticipatory in nature.

In September 2024, the UN organised a Summit for the Future in New York. As a part of the summit, the Pact for Future document was published, which touched upon digital inclusivity development and the Digital Compact, among other issues. It connected the development of digital resources and their optimum use with the Sustainable Development Goals. Yet, however, while the document was a welcome step, it lacked direct actionable points for the member state to work upon. Another similar document which also connects SDGs with Just AI is the UN 2.0 Policy Brief, that goes a step further to recognises the use of data for behavioural science in developmental aid works. Other UN publications such as International Data Governance : Pathways to Progress (2023) and The Governing AI for Humanity: Final Report recognise important values such as data accountability and responsibility, but fail to provide much beyond that.

While such documents are indeed a positive step towards framing the narrative of AI regulation, the need of the hour is an implementable perspective on the issue. The existing AI frameworks are limited to their national (or, at best, regional) jurisdictions. However, given the high portability of data and impact of biases in AI and its data sets, it is necessary that the world joins hands to ensure privacy and human rights in the face of rising AI penetration around us. This is especially true for developmental and humanitarian aid, as the decisions that technology could shape in distributing aid and resources could mean grave consequences for millions in deprivation around the world.

The UN has in the past produced legally binding agreements that have been ratified by member countries and are followed upon.  The imperative for a similar initiative vis-à-vis AI stands strong and must be pursued actively. In 2024, the UN launched consultations to inform AI governance. With technology advancing by the hour, what remains to be seen is how effective, and how soon, such consultations lead to a technologically inclusive future for all.   

Nandini Parashar is a Research Intern at the Digital Empowerment Foundation