A United Nations scientific panel says countries should not wait for scientists to fully understand how advanced AI systems could cause harm before putting stronger safeguards in place, arguing that the risks justify action now. The finding comes in the first thematic brief from the Independent International Scientific Panel on AI, the global scientific body on artificial intelligence established by the UN last year.
The panel is calling for significantly more attention and resources to be directed at managing emerging risks from advanced AI, along with tighter international coordination on safety and accountability. It acknowledges that individual countries are taking different legal approaches to the technology, but says that divergence does not remove the need for collective action.
Central to the panel's argument is the precautionary principle, which it says was designed for exactly the kind of problem AI presents. The panel describes loss-of-control risk as a case where potential harm may be catastrophic or irreversible even while its likelihood remains scientifically uncertain. Under the panel's reasoning, the world does not need to establish precisely how or why such incidents occur before acting to prevent them.
The precautionary principle was first enshrined in the 1992 UN Rio Declaration on Environment and Development, which held that scientific uncertainty is no excuse for delaying measures against potentially serious or irreversible harm. It has since become influential in environmental and public health policy, particularly in the European Union.
The brief lands during a busy week for AI diplomacy. World leaders are gathering in New York for the UN General Assembly, and Beijing and Washington are preparing to hold talks on AI. UN secretary general António Guterres called on governments last week to cooperate on addressing the threats posed by AI, warning that the world cannot afford a race to the bottom on AI safety.
The report also follows a run of documented AI security incidents. Since the Hugging Face hack was first reported, incidents have been recorded at OpenAI, Anthropic, Google, and Meta, including hacks on real-world targets and swarms of agents taking over online messaging boards.
For US policymakers, the panel's position adds pressure to an already crowded agenda that includes the UN General Assembly session and the upcoming US-China talks, as well as the domestic debate over how far AI safety rules should go.
More AI news from TechManNews.
