UN Scientific Panel Releases Preliminary Report on AI Governance

The United Nations’ Independent International Scientific Panel on Artificial Intelligence released its preliminary report on Wednesday, days before the Global Dialogue on AI Governance in Geneva, Switzerland, next week. Titled ‘Preliminary Report of the Independent International Scientific Panel on AI: Evidence-based assessment of opportunities, risks and impacts of AI’, the report presents an independent scientific assessment of AI capabilities, opportunities, and emerging risks, whilst providing a shared evidence base to help UN member states navigate rapid technological developments. “This collaborative effort to build a shared understanding of AI comes at a crucial turning point,” the panel stated in a press release. Governments around the world are making consequential decisions about AI amid great uncertainty, with sources of evidence and perspectives shifting very rapidly, often contradicting each other, and not always reflecting local conditions. As AI capabilities continue to increase, so do the risks and consequences of decisions taken worldwide, which is the primary challenge the panel seeks to address, according to the press release. In the preliminary report, the panel outlined its findings across seven key domains: AI science, progress, and its development direction; societal applications, including in science, health, education, and agriculture; economic implications; security, systems, and environmental implications; human rights, information, and democracy; cultural and individual development, autonomy, and child safety; and management, governance, and reliability. The panel identified a critical evidence-based challenge for decision-makers globally: policymakers need scientific evidence to regulate AI effectively, but by the time that evidence becomes clear, it may be too late to act. “Whether AI’s potential is realised fairly will depend on informed decisions made collectively by nations and the shared scientific foundation guiding them—the very evidence base this panel’s work aims to provide,” the panel said. Panel co-chair Yoshua Bengio noted that AI capabilities are evolving faster than scientific understanding and governments’ ability to adapt. “With mounting evidence of deceptive AI behaviours, current science cannot guarantee that as these capabilities continue to grow, AI will not cause catastrophic harm, either autonomously or through malicious users. To act effectively, global policymakers must understand these systems,” Bengio said. Maria Ressa, another co-chair of the panel, said the technology is transformative but warned that if the world continues on this trajectory, humanity will not be able to realise its promised benefits. “The risks are too high, and the forces driving AI development are not the forces that will deliver its benefits,” Ressa said. UN Secretary-General Antonio Guterres said the panel’s report presents independent science, gathered from every region, and made available to every government. “Its message is clear: the potential is enormous, but the risks are real, and the costs of delay are mounting. I urge all leaders to use this shared evidence to act together, and without delay,” Guterres said.