Written by Professor Yuval Shany, Senior Inaugural Fellow of the Accelerator Fellowship Programme.
On 6 October 2026, the Human Rights Council (HRC) adopted without a vote Resolution 63/15 on the implications of artificial intelligence on the promotion and protection of human rights. While this is not the first HRC resolution on the impact of new digital technologies, including AI, on human rights (resolutions to this effect were adopted in 2019, 2021, 2023 and 2025), it is the first detailed resolution dedicated only to the human rights impacts of AI systems. Hence, it signifies the growing interest of states and inter-governmental organizations in specifying human rights standards relevant for addressing the challenges posed by AI systems. Incidentally, Resolution 63/15 was adopted one day after the Oxford Institute for Ethics in AI held its first Summit on AI and human rights in which a number of prominent speakers called to further foreground human rights in AI governance.
While the Resolution is a welcome development, it falls short in several ways. This blog summarizes the text of the Resolution and then offers a commentary below.
The Resolution
The Resolution, which was co-sponsored by 13 states from four different UN regions, is roughly divided into two parts – a lengthy preambular section and 23 operative paragraphs. The preamble – which itself has three segments – starts with identifying a number of UN resolutions, treaties and standard-setting bodies, which contribute to the development of human rights for AI systems: These include general human rights standards like those found in the International Covenant on Civil and Political Rights and the International Covenant on Economic, Social and Cultural Rights and the United Nations Guiding Principles on Business and Human Rights; language on human rights found in regulatory initiatives such as the Global Digital Compact and the preliminary report of the Independent International Scientific Panel on Artificial Intelligence; and past HRC resolutions and UNESCO Recommendations. The preambular paragraphs also mention in this regard the 2026 Global Dialogue on Artificial Intelligence Governance and the work of the Office of the High Commissioner for Human Rights (OHCHR) and the HRC Advisory Committee and special procedures and mechanisms.
The preamble then refers to a host of potential human rights impacts that may derive from the development and use of AI systems. These include positive impacts in areas such as health, education, research, access to information and development, but also a host of potential negative impacts, which could undermine the full enjoyment of human rights. Among the rights which the Resolution mentions explicitly or implicitly as being at risk of being adversely affected by AI systems, one can find non-discrimination (including profiling), the right to privacy, freedom of opinion and expression (including access to information), the right to participation in public affairs, freedom from violence and harm, and decent labor standards. Other potential impacts noted involve the environment (affecting the right to a clean, healthy and sustainable environment), digital divides and gaps within and across countries, improvements in the capacity to deliver services at scale and in economic and social opportunities, and negative effects on persons in situations of vulnerability and marginalization (especially children and victims of racism) and on “low-resources languages” and cultural diversity.
Finally, the preamble maintains the need to ensure “that human dignity and human agency should remain central to the design, development, deployment, use and governance of artificial intelligence” and to promote “transparency and explainability, accountability, accessibility, appropriate safeguards, meaningful human oversight and access to effective remedies in addressing its human rights implications”. It calls for human rights due diligence and impact assessment to assist states and businesses in identifying and addressing adverse impacts of AI systems, underscores the importance of participation and dialogue by multiple stakeholders and calls for capacity development, knowledge-sharing, technology transfer and access to open models that would ensure “full, equal and meaningful participation” by developing countries in AI-related discussions and processes.
The first set of operative paragraphs reaffirms the potential of AI systems to generate both positive and negative impacts on human rights, and the applicability of international law to AI design, development, deployment, use and governance. This may include international human rights law (including principles governing business and human rights), international humanitarian law and the principles of the UN Charter. States are then called upon to consider the implications of AI systems on human rights when developing their own regulatory frameworks and to take appropriate measures to address the negative human rights impacts of these systems. Related AI governance measures should factor in different national contexts and capacities, promote participation of other states in AI regulation, ensure access to information about AI systems and facilitate partnerships between governments and other stakeholders.
Paragraph 7 – which is one of the key paragraphs of the Resolution – sets out a list of principles that should form part of AI governance:
[E]ncourages States to adopt and promote approaches to artificial intelligence that respect human dignity and human agency and to uphold the principles of equality and non-discrimination, privacy, accessibility, transparency and explainability, accountability, sustainability, safety, human oversight and access to effective remedies in accordance with their obligations under international human rights law and, where applicable, under international humanitarian law
The following paragraphs also encourage states to carry out human rights due diligence and impact assessment, and to recognize the need for harm prevention, especially when systems operating outside human control could be involved.
Paragraphs 10-13 of the Resolution draw attention to specific human rights implications of AI systems – the need to obtain human oversight, human review, transparency and accountability when AI systems impact the enjoyment of human rights, the need to protect and preserve a balance between freedom of opinion, expression (including access to information) and privacy, the disparate impact of AI systems on persons and groups in situations of vulnerability or marginalization and the need to facilitate their participation in relevant policy-making processes, and the responsibility of businesses involved in the life cycle of AI systems to respect human rights and to provide for or cooperate in remediation for impacts caused or contributed to.
Paragraphs 14-15 call for promoting digital literacy, international cooperation and capacity-building. Paragraph 16 encourages, in this vein, states, the UN and other stakeholders “to promote digital public goods, digital public infrastructure and open artificial intelligence models, data sets and tools, accompanied by adequate human rights safeguards, as a means of enabling countries with limited resources, to benefit from these technologies”; and paragraph 17 encourages “greater coherence and complementarity among relevant international, regional, national and local initiatives on artificial intelligence, while recognizing the important role of the United Nations in promoting inclusive international dialogue and cooperation in this field”.
In order to further develop the role of the UN in human rights-centric AI governance, the final paragraphs of the Resolution call on the OHCHR to contribute to UN discussions and processes concerning AI, support (alongside other UN agencies) capacity-building and technical development relating to human rights implications of AI systems, convene a future HRC panel discussion on the topic and prepare a report on it.
Commentary
While the Resolution builds on previous HRC resolutions on the impact of new and emerging digital technologies on human rights, its special focus on the positive and negative impacts of AI systems on human rights and the need to recenter human rights in national and international AI regulatory efforts should be welcomed. It is particularly useful that the Council is drawing attention in this context to some unique features of AI systems – the potential for manipulation, bias, privacy infringement, lack of human oversight, lack of transparency and accountability and ineffective remedies – which could have dramatic human rights implications.
Other positive features are the emphasis of the Resolution on procedural steps – due diligence, impact assessment and remediation measures – which ought to be taken by both states and business entities, the need to promote inclusive and multi-stakeholder AI human rights governance and the associated need to empower vulnerable and marginalized persons and groups and to develop the capacity of developing states.
The Resolution does fall short in a number of respects, however. It reaffirms the applicability of existing international human rights law norms (and of norms associated with other international law regimes – international humanitarian law and the UN Charter), but does not take a clear stance on whether additional legal developments – for instance, reinterpretation of existing human rights, institutional reform and recognition of new human rights and remedies – are needed. Furthermore, the measures it recommends the OHCHR to take – to support other UN processes, capacity development and to foster further discussion on the topic of AI and human rights – appear to be underwhelming in their scope and nature, and somewhat at odds with the sense of global crisis and of the inadequacy of existing norms and institutions that informed many of the discussions held earlier this month at the Oxford AI and human rights Summit. Rather, the HRC appears to largely follow its traditional “normative equivalency” approach to the application of human rights to the challenges posed by new and emerging digital technology – that is, it continues to operate on the assumption that existing human rights law norms and institutions are agile and flexible enough to accommodate major changes in the environment in which human rights are enjoyed that have been brought about by dramatic technological developments.
As I have explained elsewhere with Daphna Dror Shpoliansky, the normative equivalency approach does not afford sufficient weight to the radical changes in the world in general, and in the world of human rights in particular, facilitated by new and emerging digital technologies, including AI technology. This might result in the shoehorning of new problems into inadequate pre-existing legal structures that were created for dealing with very different types of challenges presented by quite different actors, often at a different scale of severity and with narrower whole-of-society implications. By contrast, acknowledging the need for legal adaptation and innovation could help stakeholders, including potential victims, to better articulate their normative expectations and demands, promote new thinking about the effectiveness of current avenues of redress and avoid over-burdening existing courts and law-applying institutions with the difficult task of developing on their own human rights law for AI systems (a task that raises both questions of available expertise – especially given the limited propensity of potential victims to litigate many AI-related issues – and democratic legitimacy).
Arguably, it would have been desirable for the HRC and other UN stakeholders to give serious consideration to the possibility of launching a standard-setting process, which will engage with different stakeholders, and develop detailed normative and institutional guidelines for adjusting international human rights law to the challenge of AI. The process which led in the early 2000s to the development of the UN Guiding Principles on Business and Human Rights could be a good model to follow in this regard. Such a process could lead to the articulation of a list of relevant human rights safeguards, which largely elaborates upon existing human rights norms, and gives specific expression to the human rights dimensions of the many important issues identified in the Resolution – transparency and accountability, human control and oversight over sensitive decisions, and concerns about manipulation at scale (for a comparable list, see my White Paper on the topic). In addition, a standard-setting process could further clarify the specific normative expectations from different actors involved in the life cycle of AI systems, including, in particular, AI companies. Finally, greater normative clarity and institutional reform could also support and facilitate the work of national legislatures and other private and public stakeholders currently confronting the challenge of adapting human rights to the unique opportunities associated with the growing reliance on AI systems (compare the recent UK Parliament joint committee report). In sum, the HRC should be commended for turning its focused gaze to the relationship between AI and human rights; still, it should consider adopting in the near future a more ambitious normative and institutional reform agenda than the one indicated in Resolution 63/15.
Suggested citation: Professor Yuval Shany, ‘Towards a global AI human rights law? The first Human Rights Council resolution on AI and human rights’, (16 October 2026) AI Ethics at Oxford Blog; https://www.oxford-aiethics.ox.ac.uk/blog/towards-global-ai-human-rights-law-first-human-rights-council-resolution-ai-and-human-rights