Generative AI creates a societal challenge: by making prediction, generation, and procedural execution abundant, it risks pushing economies toward automation-first configurations in which human involvement becomes peripheral, symbolic, or concentrated among already-expert actors. The danger is not only job displacement, but erosion of the expertise pipeline: firms can rationally reduce entry-level training and rely on AI-assisted short-term performance, while society still depends on the slow reproduction of human judgment, accountability, and tacit expertise. The Hybrid Intelligence (HI) Manifesto series proposes a proactive response: a human-centered AI transition in which machine prediction and human judgment are structurally combined, and markets, organizations, and institutions are shaped to reward human–AI complementarity rather than pure automation. This paper develops education as the capability infrastructure of that transition: the institution through which society reproduces the hybrid capabilities needed to shape AI-mediated futures. We present the HI Educational Transformation Playbook, a four-track architecture for moving education from reactive adaptation to societal stewardship. First, mission and culture are reframed: educational institutions become proactive societal actors that cultivate citizens, professionals, and institutional entrepreneurs capable of designing, evaluating, and governing human–AI collaboration. This is operationalized through the prediction–judgment frontier, a teachable procedure for identifying where AI prediction is appropriate, where human judgment is required, and where accountable ownership must remain human. Second, the HI mindset is developed through authorship-preserving collaboration and seamful system-shaping, operationalized by FERC (Frame–Explore–Refine–Commit), a governed interaction cycle grounded in Deweyan inquiry, Vygotskian internalization, and self-regulated learning. Third, educational practice is redesigned around explicit prediction points and judgment points, illustrated through bot co-design, stakeholder simulation, flawed-bot repair, and workflow re-engineering. Fourth, learning objectives shift from task execution toward judgment-centered augmentation: framing, validating, refining, and responsibly finalizing AI-supported work. We illustrate the playbook through two formative enactments: an institution-wide transformation of a Danish business and vocational college and a planned extension into physical making at ETH Zürich’s maker space. We conclude that education is not simply where the AI transition must be managed, but one of the institutions through which society learns to steer it.
Research Article — 10 July 2026
The HI Manifesto for Education
Institutionalizing Societal Stewardship in the Age of AI
Authors: Jacob Sherson, Blerim Emruli, Rigmor Lyck Hansen, Sanna Järvelä, Yoed N. Kenett, Frederik Brosbøl Kjeldsen, Camila Kølsen Petersen, Izabela Lebuda, Yaoli Mao, Yishay Mor, Andy Nguyen, Janet Rafner, Adam Vigdor Gordon, Seyedahmad Rahimi, Matthias Söllner, Ana Alina Tudoran, Dominik Dellermann, Steve DiPaola, Florent Vinchon, Roni Reiter-Palmon, Jens Gamauf Madsen, Oday Darwich
