UN unanimously adopts ambitious AI resolution, sans teeth
'Safe, secure and trustworthy' AI a must, says document, but nothing in it ensures anyone plays along
The United Nations has unanimously adopted a resolution aimed at establishing international AI development standards.
Introduced earlier this month by the United States, the resolution, which passed the General Assembly without a vote, was backed by more than 120 other UN members, including China. The resolution's language calls for development of "safe, secure and trustworthy" AI systems that "have the potential to accelerate and enable progress toward [UN] sustainable development goals."
"[AI] has the potential to unlock more progress, on more issues, for more people," US ambassador to the UN Linda Thomas-Greenfield said in the General Assembly yesterday. "But in order to mitigate risks for communities across the globe, we must approach this technology as a global community."
"Today, all 193 members of the UN General Assembly spoke with one voice – and together, adopted a resolution on artificial intelligence for sustainable development," Thomas-Greenfield added.
The resolution, which states it only addresses AI in the "non-military domain," also calls for establishment of a global consensus on what makes a safe, secure and trustworthy AI as well as the creation of international standards to that end.
Additionally, the resolution asks member states that are further along in AI development to ensure they work with developing countries to ensure they aren't left behind as AI technology proliferates. The UN said it is urging members "to cooperate with and support developing countries so they can benefit from inclusive and equitable access, close the digital divide, and increase digital literacy."
More like guidelines
Crucial to note, the resolution is filled with language like "encourages" and "calls upon" without anything definitively enforceable, making clear the matter is non binding. The lofty ideals of the resolution, therefore, are just that: Ambitions.
Rules defining what AI can and can't do, it's worth pointing out, are scarce in the US, which introduced the UN resolution itself.
One attempt at creating a federal-level regulatory agency able to enforce proper use of AI in the US has stalled, while other efforts have been fought by politicians. The effectiveness of state-level attempts at regulating AI, meanwhile, has been questioned.
- Damn, even the Pope thinks AI and autonomous weapons need reining in
- Nations agree to curb enthusiasm for military AI before it destroys the world
- World leaders ink AI safety pacts while Musk and Sunak engage in awkward bromance
- UK and US lead international efforts to raise AI security standards
China and the European Union, on the other hand, are further ahead with their attempts to regulate AI, with China setting restrictions on the tech last year, and the EU passing the world's first legislation designed to address AI risks earlier this month.
This isn't the first UN initiative on AI either, the resolution notes. It aims to support work already being done by the International Telecommunication Union's Artificial Intelligence for Good program, the High Commissioner for Human Rights AI-related work and other programs.
"We intend for [the resolution] to complement future UN initiatives, including negotiations toward a global digital compact and the work of the Secretary-General's high-level advisory body on artificial intelligence," Thomas-Greenfield said. ®