THE rapid advancement of artificial intelligence (AI) presents vast opportunities for economic and social development but also significant risks that are growing faster than the world’s ability to regulate or fully understand the technology, according to a United Nations-backed scientific assessment.
In its first report, produced by an independent panel of 40 leading scientists and experts from around the world, the UN said policymakers urgently need stronger scientific evidence to govern AI systems whose capabilities are evolving at a pace that outstrips regulatory frameworks.
Reuters cited on Wednesday that the report, which will be presented to governments at the inaugural UN Global Dialogue on AI governance in Geneva on 6–7 July, is the first global independent scientific assessment of artificial intelligence. A more comprehensive version is expected next year.
Panel members, serving a three-year independent term free from government, institutional or corporate influence, warned that current tools for controlling highly autonomous AI systems remain limited.
Panel co-chair Yoshua Bengio highlighted growing evidence of deceptive behaviour in advanced AI systems, cautioning that science cannot guarantee AI will be prevented from causing catastrophic harm, whether through autonomous actions or malicious use as capabilities increase.
“The potential benefits of AI are enormous,” the report stated. “The rapid, unchecked deployment of the technology at scale also presents considerable risks, including harms to the mental health of users, potential use as a destructive tool, impacts on social, economic and environmental systems, and challenges associated with controlling the technology.”
The report noted that AI adoption has expanded rapidly but unevenly across countries and sectors, with more than one billion people now using conversational AI weekly, while many developing countries continue to lag behind.
It also highlighted stark concentration in AI development capacity, with the United States accounting for around 75 per cent of computing power among the world’s top 500 AI supercomputers, compared with about 15 per cent for China.
The panel further warned that linguistic limitations in current AI systems could deepen global inequality, noting that while more than 7,000 languages are spoken worldwide, models are trained on only a small fraction, leading to errors in machine translation that can affect critical areas such as healthcare.
Among the risks identified were the proliferation of AI-generated child sexual abuse material, deepfake-enabled sexual violence, and the broader erosion of information integrity through the mass production of persuasive content, which could undermine public trust, social cohesion and democratic processes.
The report also cautioned that most countries, including advanced economies, lack the technical expertise required to properly evaluate innovative AI systems or meaningfully participate in their governance. - July 1, 2026
.jpeg)
