UN Panel Warns AI Is Outpacing Oversight

Artificial intelligence is advancing faster than scientific research and government policymaking can keep up, raising concerns that the technology could cause serious harm in the future without adequate safeguards, a United Nations-backed independent panel of experts warned on Wednesday.

In a preliminary report, the UN's Independent International Scientific Panel on Artificial Intelligence said policymakers are increasingly confronting a complex reality. Effective AI governance depends on reliable scientific evidence, yet research is struggling to keep pace with the technology's rapid development.

"AI capabilities ​are outpacing both scientific understanding and governments’ ability to adapt," said Yoshua Bengio, co-chair of the 40-member international panel.

Bengio said evidence of deceptive behavior by AI systems is growing. He warned that current scientific knowledge cannot guarantee that increasingly powerful AI systems will not cause severe harm, either independently or through misuse by individuals with malicious intent.

The report describes itself as the first independent global assessment of both the risks and opportunities presented by artificial intelligence. It aims to provide governments with the latest scientific evaluations to support decision-making on a technology that is evolving at an extraordinary pace.

According to the report, so-called ‘agentic AI’ systems capable of autonomously carrying out real-world tasks are likely to become more widespread in the near term. However, constraints related to energy supply and shortages of high-quality data could slow the pace of development.

Looking further ahead, the panel projects that self-improving AI systems will become more deeply integrated across economic sectors. The report also forecasts that AI will increasingly converge with advanced technologies such as quantum computing and biotechnology, creating new opportunities and capabilities.

 Source: Reuter (adapted)