The United Nations’ Evolving Role in Global AI Governance
In our last post, we posed a question: Is Global AI Governance Possible? and gave a few examples of notable industry organizations that currently provide such governance on a global scale. Although AI doesn’t easily fit within any of those industry paradigms – it’s unique on many fronts – the basic intents of a governing body still apply.
rules + standards + safeguards + enforcement + verification + notification
Which brings us to the primary organization that is attempting to address these pressing issues regarding AI and provide much-needed governance: the United Nations. The UN is no stranger to information technology and its impact on the world. Their efforts, spanning the past 20 years, have been significant.
- 2006 – Internet Governance Forum (IGF): Born out of the World Summit on the Information Society (WSIS) Tunis Agenda, the IGF was established as a multi-stakeholder platform for governments, technology companies, civil society, and academia to discuss internet policy issues.
- 2017 – AI for Good Global Summit: Created as the main operational platform within the UN to connect various AI innovators with practical applications for the Sustainable Development Goals (SDGs).
- 2018 – High-Level Panel on Digital Cooperation: Appointed by Secretary-General António Guterres and issued the Age of Digital Interdependence report the following year, advocating for human-rights-based digital architecture.
- 2020 – The UN Secretary-General’s Roadmap for Digital Cooperation: Outlined action items including universal connectivity, protecting human rights online, and international coordination on AI governance.
- 2021 – UNESCO Recommendation on the Ethics of Artificial Intelligence: Adopted by 193 member states, this was the first global normative framework on AI ethics, establishing core principles such as transparency, accountability, and the protection of privacy.
- 2023 – UN High-Level Advisory Body on Artificial Intelligence: Convened to assess risks, opportunities, and international governance mechanisms for AI. Its final report, Governing AI for Humanity, published in 2024, recommended establishing an international scientific panel on AI and an annual global AI policy dialogue.
- 2024 (March) – First UN General Assembly Resolution on AI: Adopted by consensus, the resolution titled “Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development” marked the first time the General Assembly collectively set global baseline guidelines for AI development.
- 2024 (September) – Global Digital Compact: Adopted as part of the Pact for the Future at the UN Summit of the Future. The Compact provides a multi-stakeholder framework for setting commitments on digital connectivity, data protection, digital public goods, and international AI governance.
- 2026 (March) – First Global Advocate for Human-centric Digital Governance – This newly created role affiliated with the United Nations Internet Governance Forum (IGF) advocates for inclusive, rights-respecting approach to internet and AI governance, aiming to make complex digital policy discussions more accessible to the general public.
The Need for AI Governance Today (and in the future)
So what are the issues that the United Nations needs to be thinking about? To answer that question it would be helpful to review some of the critical incidents that have occurred this year.
Reported incidents involving AI agents accessing external systems in 2026
| Date | Organization / system | Incident summary | Source |
|---|---|---|---|
| July 2026 | OpenAI agents → Hugging Face | Agents reportedly escaped a test environment by exploiting a vulnerability, then used credentials and additional vulnerabilities to access Hugging Face infrastructure. | OpenAI — incident report |
| July 2026 | OpenAI internal research systems | In the broader evaluation episode, agents reportedly obtained unauthorized internet access and reached administrative access on a research cluster. | OpenAI — follow-up report |
| 2026, disclosed July | Anthropic Claude Opus 4.7 → unnamed company | During a cybersecurity evaluation, the model reportedly exploited vulnerabilities and obtained credentials and access to a database containing production data. | Anthropic — incident assessment |
| 2026, disclosed September | Anthropic Claude Mythos 5 → software supply chain | During a test, the model reportedly published a malicious Python package to a public registry. Automated security systems removed it. | Anthropic — incident assessment |
| 2026, disclosed July | Anthropic research model → unnamed company | An evaluation agent reportedly scanned thousands of internet targets and compromised an exposed application. The account says the agent later recognized the system was real and stopped. | Anthropic — incident assessment |
| January 2026, disclosed September | Anthropic Claude Opus 4.6 → real system | Anthropic reportedly identified another evaluation incident involving access to a real third-party system, with fewer public details available. | Anthropic — incident assessment |
| May 2026, disclosed September | Google Gemini → three companies | During security tests, Gemini reportedly accessed three companies’ systems, including through a guessed password and publicly exposed credentials. | Alexander Martin, The Record |
| July 2026 | Frontier models → real-world targets | In permissive cybersecurity tests, agents reportedly attempted unsanctioned actions against real organizations, including trying to insert malicious code into an open-source project. The reported attempts were unsuccessful. | UK AI Security Institute — incident report |
| June 2026, disclosed September | OpenAI agent → Australian government health portal | An agent reportedly accessed non-public files on a Medicare statistics portal and wrote files to an internal server. The government reportedly found no evidence that personal information had been accessed. | Alexander Martin, The Record |
| July 2026 | OpenAI agents → other online services | OpenAI reportedly said agents used exposed credentials to access four accounts across four services. The reported access varied, and no broader impact was identified. | OpenAI — follow-up report |
How can the United Nations shape our global AI future?
Humanity sits at a critical crossroads, wondering if our future will be bright and inspiring, or become a darker, more threatening type of existence. Is it a choice between promise and peril? And to what extent can the United Nations further influence AI’s development and deployment?
Image: Mark Lovett / AI GAPS — Generated with ChatGPT Images 2.5
The UN can play an important role by helping countries agree on a common framework for testing frontier models, as well as reporting, investigating and learning from significant AI incidents. Lessons learned can then be used to develop internationally recognized standards for AI safety, oversight and accountability.
This is not about the UN becoming some sort of global AI police force, but instead would take advantage of their unique ability to convene countries, establish shared expectations, draw on independent science and internal evaluators in order to coordinate international cooperation.
As we know, AI is increasing its capabilities at a rate that exceeds any technology the world has developed in the past. And the fundamental problem we’re facing at the moment is that our ability to deal with AI governance, alignment, policy, and safety issues are not keeping pace. So we need all the help we can get.