top of page
Search

Opinion: The missing conversation at the first UN Global Dialogue on AI Governance

Writer: Planetary Collaboration
Planetary Collaboration
Aug 3
4 min read

The first UN Global Dialogue on AI Governance was, in one aspect, disappointing. With the exception of UN Secretary-General António Guterres Guterres and the UN High Commissioner for Human Rights Volker Türk, no one addressed the use of AI in military and security applications.

That omission should not overshadow the Dialogue's achievements. It united 191 nations and gave an equal democratic seat to voice their needs. On the other hand, most speakers were addressing essentially the same points described in the Preliminary Report of the Independent International Scientific Panel on AI, which was launched a few days before the Dialogue. 

Governing AI reactively

This was a high-level dialogue where heads of state, ministers, and senior representatives shared their priorities for AI governance. The issues they raised were important, but they felt like yesterday's agenda. Much of the discussion appeared driven by the consequences of Silicon Valley's "move fast and break things" approach, focusing on repairing damage that has already occurred instead of anticipating the next generation of risks.

This was addressed by Maria Ressa, Nobel Peace Prize laureate and co-chair of the Scientific Panel on AI. "The risks are real. I watched this a decade ago with machine learning and AI on social media. It promised to connect us, and instead it pulled apart our shared reality... until the damage was done and it was too late to act... Please, we cannot do that again", she reinforced.

AI is not only software

AI is increasingly embodied in robots, surveillance systems, autonomous vehicles and, critically, autonomous weapons. Ignoring this dimension risks overlooking one of the most significant governance challenges of our time.

Among all the speakers, only Mr Guterres and Mr Türk explicitly referred to threats to human security, including military applications of AI, autonomous weapons and the broader implications of increasingly autonomous systems. These issues deserve far greater attention in global AI governance discussions.

There was a reason for this omission. As Mr Guterres himself noted, the Global Dialogue focuses on "civilian AI." But he immediately challenged that distinction:

"The global dialogue is about civilian AI, but AI does not respect that line. The same models and chips have moved into the battlefields. My main concern is with lethal autonomous weapon systems. Let's call it what they are: killer robots."

That statement goes to the heart of the issue.

Are we creating a new governance silo?

If, according to Mr Guterres,  AI does not respect the line between civilian and military use, why should this distinction occur at the UN Global Dialogue on AI?

The scope of mandate of both the Global Dialogue on AI Governance and the Scientific Panel on AI are described in the General Assembly Resolution 79/325, which explicitly limits the activities to the non-military domain:

"activities of the Panel and the Dialogue are limited to the non-military domain and do not refer to artificial intelligence for military purposes".

The UN frequently calls for breaking down silos to address complex global challenges. Yet by separating civilian AI from its military applications, it risks creating another silo that prevents a comprehensive discussion of AI's societal impacts.

Effective AI governance requires looking at the entire ecosystem, including the technologies that directly affect human security, peace, and international stability.

Main topics addressed at the Dialogue on AI

The conversation on Agentic AI, with human oversight and accountability, seems to be in a good moment, though. This technology is advancing fast but still in its early steps, as it still requires a certain level of specialized engagement to build agentic AI systems. This, however, will not be the reality soon, as technology advances and entire agentic systems will be created with a simple prompt.

Other topics include:


  • International cooperation and Human Rights as the foundation for AI governance were recurring themes.

  • Speakers also emphasized the need for civil society participation. The French representative mentioned having people not as subjects but as co-creators.

  • The protection of children and women. The President of the General Assembly, Annalena Baerbock, noted that a reported 99% of deepfakes are sexual in nature and 96% target women and girls.

  • Multilingualism was another interesting point raised by some participants. The Government of Guatemala talked about the recognition of cultural and linguistic diversity, and the need to integrate indigenous languages and respect community and ancestral knowledge. To exemplify the relevance of this topic, the co-chair Ressa reported that "Intgna, spoken by seven to 9 million people in Eritrea and northern Ethiopia, machine translation turned smallpox into syphilis, gonorrhea into diabetes... That can be life-threatening, someone who can't get correct medical information in their own language."

  • Youth inclusion was also a concern expressed by member states. During audience intervention, Luísa Franco Machado said that even though there are talks about youth inclusion, there were barely any youth in the room. She asked people with less than 30 years old to stand up, which only a handful of people did.


These are all essential elements of responsible AI governance, but the technical aspects were largely centred on AI as software, particularly large language models.


Note: This is an opinion piece intended to contribute constructively to the ongoing work of the United Nations on AI governance. The First UN Global Dialogue on AI Governance was an exceptionally rich and valuable forum, bringing together 191 Member States and a wide range of stakeholders. It would not have been possible to address every important issue discussed during two days in a single article. The observations above are offered in the spirit of strengthening future dialogues and broadening the conversation on emerging AI governance challenges.


Sources



Comments

Rue Jean-Dassier 7

1201, Genève

Switzerland

bottom of page