wsf-policy-brief-2026-002-before_the_warning

Published by the World Sustainability Forum in October 2026, this co-authored policy brief examines how institutions can recognise emerging risks when indicators and classifications no longer capture changing interactions among technological, economic, ecological and social processes. It proposes morphological sensing as an institutional capacity that investigates anomalous evidence without automatically converting novelty into a prediction or warning. The authors explain how machine learning, change-point detection, graph analysis and language systems can assist inquiry while remaining subject to independent verification, human judgement and alternative interpretations. They identify five evaluative tests covering unfamiliar evidence, independent validation, competing representations, revisable categories and institutional responsiveness. Drawing on United Nations AI-governance initiatives, the brief recommends a limited comparative pilot to assess whether these practices improve anticipatory analysis without weakening accountability, human rights or public oversight.