The United Nations Draws a Line on Artificial Intelligence

Two resolutions and one warning

More than sixty armed conflicts involving at least one State are being fought around the world today, a level not seen since the end of the Second World War. Into this troubled landscape, artificial intelligence has arrived not as a distant prospect but as a working tool: it sorts, targets, persuades and, in an increasing number of documented cases, it kills. On 6 October 2026, at the close of its sixty third session, the United Nations Human Rights Council adopted two resolutions addressing the human rights dimension of this transformation. Days later, in New York, the United Nations High Commissioner for Human Rights, Volker Türk, told the Third Committee of the General Assembly that the world had reached, in his words, the final level of red alert.

Two texts adopted by consensus

The first resolution, A/HRC/RES/63/13, carries the title “Strengthening international cooperation for promoting and protecting the rights of women and girls in the context of digital technologies”. The second, A/HRC/RES/63/15, is entitled “Implications of artificial intelligence for the promotion and protection of human rights”. Both were adopted without a vote at the forty first meeting of the sixty third session of the Council, which ran from 7 September to 7 October 2026.

The absence of a recorded vote matters. It indicates that, at least at the level of principle, the membership of the Council was able to converge on a common language about artificial intelligence and human rights. That convergence should not be mistaken for unanimity of practice. The text of resolution 63/15 itself acknowledges that the adoption of artificial intelligence has accelerated broadly and unevenly, and that the wealth it creates is highly concentrated. Still, a shared normative frame is a precondition for any effective accountability, and the Council has now supplied one.

Resolution 63/15: artificial intelligence under human rights law

Resolution 63/15 is the more general of the two texts, and it reads as a systematic inventory of what artificial intelligence can do to people, for better and for worse.

On the positive side, the Council recognises that artificial intelligence, when adequate human rights safeguards are in place, can contribute to the promotion and protection of human rights and fundamental freedoms, including by facilitating access to information, education, healthcare and public services. This recognition is not rhetorical decoration. It matters because it anchors the resolution in a balanced assessment rather than in technophobia, and it gives developing countries a legitimate claim to share in the benefits of the technology.

The concerns listed are longer and more concrete. The Council recognises with concern that artificial intelligence systems may reproduce and exacerbate patterns of discrimination, structural inequalities, marginalisation, stereotypes, biases and exclusion. It points to infringements of the right to privacy, arbitrary and unlawful surveillance, threats to information integrity, large scale persuasion, distortion of public debate and the erosion of trust in democratic processes. It also notes environmental harms across the life cycle of the technology, from energy and water consumption to the extraction of critical minerals and electronic waste.

For religious minorities and those who defend freedom of religion or belief, one preambular paragraph deserves particular attention. The Council expresses concern that artificial intelligence systems, including those relying on automated profiling, biometric technologies, predictive analytics and algorithmic decision making, may reproduce, reinforce or amplify existing patterns of racism, racial discrimination, xenophobia and other related intolerance, including intolerance and discrimination based on religion or belief. This is a direct acknowledgment, in a consensus text, that algorithmic systems can become instruments of religious discrimination.

The operative paragraphs translate these concerns into instructions. States are called upon to consider the implications of artificial intelligence for the enjoyment of all human rights when developing and implementing legislative, regulatory and policy frameworks. Paragraph 8 is among the strongest: the Council recognises the necessity to prevent harm and to refrain from, or cease, the use of artificial intelligence applications that are impossible to operate in compliance with international human rights law or that pose undue risks to the enjoyment of human rights, including systems whose capabilities could reasonably be expected to cause the loss of human control. Paragraph 10 requires, where artificial intelligence plays a material role in decisions affecting human rights, mechanisms ensuring meaningful human oversight, access to information about such decisions, human review and effective remedies.

The resolution also addresses the private sector directly, reminding business enterprises involved at every stage of artificial intelligence systems of their responsibility to respect human rights under the Guiding Principles on Business and Human Rights, including through human rights due diligence. It keeps the process moving: a panel discussion on artificial intelligence capacity building and the digital divide will be held at the sixty fifth session of the Council, and a report on its theme is due at the sixty sixth session, informed by consultations including at the margins of the Second Global Dialogue on Artificial Intelligence Governance, to be held in New York in May 2027.

Resolution 63/13: the gendered dimension

Resolution 63/13 addresses a gap that general frameworks tend to leave open: the specific ways in which digital technologies, and artificial intelligence in particular, affect women and girls. The Council recognises that discriminatory design, algorithmic and other forms of bias, the gender digital divide within and among countries and the lack of representativeness of data may reproduce or amplify existing gender inequality and stereotypes. It notes that violations and abuses of the rights of women and girls online are a global concern that may deter them from using information and communications technology, thereby deepening the divide.

The operative part urges States to respect, protect and fulfil the human rights of women and girls throughout the research, design, development, deployment and use of digital technologies. It calls for gender responsive measures to bridge the gender digital divide, for effective measures against technology facilitated gender based violence, and for the identification and mitigation of algorithmic bias in the training of artificial intelligence systems. The resolution ties the Council to the wider architecture that has emerged since 2025: the Independent International Scientific Panel on Artificial Intelligence and the Global Dialogue on Artificial Intelligence Governance, both established by the General Assembly in its resolution 79/325 of 26 August 2025. An annual panel discussion on the human rights of women and girls in the context of digital technologies will be held at the sixty eighth session, with a report to follow at the seventy first.

The final level of red alert in New York

A week after the resolutions were adopted, the Third Committee of the General Assembly opened its annual dialogue with the High Commissioner for Human Rights in New York. The exchange, held on 14 October 2026, was dominated by the same subject: artificial intelligence and armed conflict.

“When you listen to tech specialists, the alarms are sounding louder and louder: we have reached the final level of red alert,” Volker Türk warned the delegates. He regretted that human rights are too often the last dimension that comes to mind for those who create new technologies, observing that with proper training in the subject, some of these tools might never have been developed. Human rights, he reminded the room, are above all normative obligations that both States and companies must respect.

On military applications, his language was stark. Deploring what he called the unprecedented risks posed by the use of artificial intelligence in warfare, he said that autonomous lethal weapons and drones are creating “a new circle of hell for civilians around the world”. He judged it urgent to design a system of artificial intelligence governance that works before it is too late. This echoed earlier warnings: in his opening address to the Human Rights Council in September 2026, he had spoken of an existential threat and of reports that fully autonomous drones killed three Ukrainians the previous month, joining calls for the prohibition of weapons that take lives without human involvement.

The High Commissioner’s alarm landed in a divided room. Several States used the dialogue to attack his Office directly, accusing it of bias and flawed methodology. Türk replied that defending human rights often means revealing uncomfortable truths. The Republic of Korea supported his call to keep human rights at the centre of global artificial intelligence governance and expressed its readiness to take part in the global dialogue in New York in May 2027. Many delegations also raised the severe underfunding of the human rights pillar, which receives only one per cent of the regular United Nations budget despite carrying the largest number of mandates.

Why this matters for freedom of religion or belief

For CAP Liberté de Conscience, these developments are not abstract. Our organisation has documented for three decades how profiling, surveillance and discriminatory data practices affect religious minorities across Europe and beyond. Resolution 63/15 explicitly names discrimination based on religion or belief among the patterns that artificial intelligence can reproduce and amplify. Automated profiling of worshippers, biometric surveillance of places of assembly, predictive analytics applied to migration and the administration of justice: each of these use cases touches directly on the freedom to manifest one’s religion or belief, alone or in community with others.

CAP LC therefore welcomes the Council’s insistence on human rights due diligence, impact assessments, transparency, meaningful human oversight and access to remedies, and its reminder to technology companies of their responsibilities under the Guiding Principles on Business and Human Rights. The Third Committee dialogue adds a second lesson: the institutions that monitor artificial intelligence and its human rights impacts, including the Office of the High Commissioner itself, are under political and financial pressure at the very moment their work becomes most necessary. The independence of these mechanisms is not an administrative detail. It is the condition for the credible documentation on which every defence of religious minorities ultimately relies. In line with the resolutions, CAP LC will continue to monitor the use of artificial intelligence in the surveillance and profiling of religious communities and to report documented cases to the relevant United Nations mechanisms.

A line drawn, and a test ahead

Two resolutions, one High Commissioner, one message. The United Nations has now said, with one voice on the Council, that artificial intelligence must operate within international human rights law, that certain applications are simply incompatible with that law, and that the human being, with human dignity and human agency, must remain at the centre of every system.

The test is implementation. Resolution 63/15 asks States to conduct due diligence and impact assessments, to guarantee oversight and remedies, and to refrain from uses that cannot be squared with human rights. Resolution 63/13 asks them to close the gender digital divide and to confront algorithmic bias. Volker Türk has gone further, urging governance before it is too late and the prohibition of weapons that take lives without human involvement. Whether the gap between the language adopted in Geneva and the practices observed in more than sixty ongoing conflicts will narrow, and whether the monitoring bodies will be given the means to do their work, will measure the credibility of the entire edifice. Civil society, for its part, will be watching closely, and reporting.

Source :