This paper presents expert perspectives derived from a visioning activity conducted by an international panel of educational researchers from 13 countries and 18 institutions, working under Thematic Working Group 5 for UNESCO’s EDUsummIT 2025. Using a futures thinking approach, the experts co-constructed six scenarios imagining AI-enriched educational environments in 2030 across primary, secondary, technical, and higher education sectors. Reflexive thematic analysis of these expert-generated artefacts identified 12 cross-cutting themes. Three meta-patterns emerged: the more-human paradox, in which AI saturation intensifies rather than displaces human pedagogical work; the agency-scaffolding tension, in which student AI agency must be carefully developed rather than assumed; and contextual variation, in which the meaning of critical AI partnership differs across educational phases. Together, these meta-patterns suggest that pre-service teacher preparation moves beyond AI tool training to embed pedagogical rehearsal of scaffolding strategies that make effective use of GenAI's capabilities. We identify three priority areas that reflect the most pressing areas for action: (a) developing critical, agentic AI partnerships, (b) orchestrating AI-enhanced learning ecosystems, and (c) building trustworthy, equitable AI infrastructure. These insights aim to inform the design of teacher education and professional learning that is responsive to rapidly evolving AI-mediated educational environments.