The rapid development of Artificial Intelligence offers great benefits to countries and societies around the world, but also carries significant risks, 40 leading scientists and experts warn in the first report of an independent United Nations scientific committee.

The report, which will be presented to governments at the inaugural session of the UN Global Dialogue on AI Governance on 6–7 July in Geneva, is the first global, independent scientific assessment of the technology. A more detailed report is expected to be published next year.
The committee is composed of experts selected from all regions of the world and operates independently of governments, institutions or companies.

The report says public policy urgently needs scientific evidence to "govern" AI, but the technology is advancing faster than scientific understanding and governments' ability to adapt. It also highlights the lack of effective methods for controlling increasingly autonomous systems.

The committee's co-chair, Yushua Benzio, has raised concerns about growing evidence of deceptive behavior by AI systems, warning that the technology could cause serious harm "both from the systems themselves and from malicious users."

The report emphasizes that, while the benefits of AI are "tremendous," its uncontrolled development could bring risks to mental health, social systems, the economy and the environment, as well as challenges in controlling the technology.

Another concern is the uneven use of AI around the world: over a billion people now use it every week, but developing countries lag behind. The US and China also dominate the global AI infrastructure, controlling the vast majority of computing power.

Experts also raise the alarm about the linguistic limitations of existing models, which have only been trained in a small fraction of the world's over 7,000 languages, which could lead to serious errors in critical fields such as medicine.

The report also warns of the increasing risk of disinformation, deepfakes and sexual abuse material generated by AI, stressing that these developments could undermine public trust and democracy.

According to the UN, most countries do not yet have the technical capacity to effectively assess or govern the most advanced artificial intelligence systems.