International governance measures are being considered in a growing world of artificial intelligence (AI). Recently, the Secretary-General of the United Nations, Antonio Guterres, endorsed a proposal put forth by prominent AI industry figures. They envision establishing a global AI watchdog modeled after the International Atomic Energy Agency.
The remarkable rise of generative AI technologies, which can generate compelling prose from simple textual prompts, provides the context for this development. For example, ChatGPT, an AI application released only six months ago, has quickly become the most widely used app ever. The public’s preoccupation with artificial intelligence is palpable, but growing concerns exist. The ability of artificial intelligence to generate deep fake images and distribute false information has sparked widespread concern.
Guterres emphasized the situation’s urgency, stating, “The alarm bells are ringing loudly over the latest form of artificial intelligence – generative AI. The warnings are loudest from the very developers who designed it. We must heed these serious warnings.”
In response, the UN Secretary-General announced his intention to launch work on a high-level AI advisory body by the end of the year. This entity will periodically evaluate AI governance measures and propose solutions to ensure their conformity with human rights, the rule of law, and the common good.
Guterres added, “I am receptive to the idea of establishing an artificial intelligence agency, drawing inspiration from the international atomic energy agency of today.” Moreover, he remarked that this model could be “very interesting” but emphasized that “only member states can create it, not the United Nations Secretariat.” The Vienna-based IAEA was founded in 1957 to promote the safe, secure, and peaceful application of nuclear technologies while monitoring potential Non-Proliferation Treaty (NPT) violations. Currently, there are 176 member states in the IAEA.
The concept of a global AI overseer has also resonated with the OpenAI creators of ChatGPT. Last month, they proposed that an organization similar to the IAEA could impose restrictions on the deployment of artificial intelligence, assure compliance with safety standards, and monitor the use of computational power.
The concept has also garnered the support of British Prime Minister Rishi Sunak, who has expressed his wish for the United Kingdom to become the global epicenter of AI safety regulation. Later this year, Britain will host a summit to discuss how coordinated international action can mitigate the risks associated with AI.
In this context, Guterres expressed his support for the upcoming summit in the United Kingdom and emphasized the need for extensive preparatory work. He also intends to appoint a scientific advisory council of AI experts and chief scientists from various UN agencies in the near future.