AI Can Revolutionise Education but Technology Is Not Enough: Human Development Meets Cultural Evolution

作者：Michael Muthukrishna, Jiner Dai, Diana Panizo Madrid, Riya Sabherwal, Karlijn Vanoppen, Hanying Yao · 发表于：Journal of Human Development and Capabilities · 年份：2025 · DOI：10.1080/19452829.2025.2517740 · 被引用次数：3 · 研究领域：Cognitive Science and Mapping、Language and cultural evolution、Cognitive Science and Education Research

Artificial Intelligence could dramatically boost educational outcomes and close gaps – but only if policymakers take a human-centred, systems-level approach to AI integration. Cultural evolution, the science of how beliefs, values, norms, technologies and institutions evolve over time, offers a framework for understanding the promises and pitfalls of different approaches to AI in education policy. Using this perspective and drawing on comparative evidence from Estonia’s successful “Tiger Leap” initiative and the failed “One Laptop Per Child” (OLPC) programme, we identify three missteps that derail national strategies: (1) techno-fix thinking, (2) weak infrastructure and teacher support and (3) lack of local adaptation. Uruguay’s Plan Ceibal is a notable exception to OLPC’s general failures, revealing why technology alone is not enough. We map AI’s headline promises – personalised tutoring, higher teacher productivity, smaller equity gaps – onto the specific capabilities each can expand, and we highlight three systemic risks: digital exclusion, algorithmic bias and widening inequalities. Synthesising these lessons, we propose a practical roadmap. AI will revolutionise education and enhance human development only insofar as it is embedded in human centred systems that grow everyone’s capabilities and freedom to learn, create and participate in society.