Educate Ventures Research: Human Intelligence First
Human Intelligence First: the Tool is the Smallest Part
Primary Source: Educate Ventures Research report page · Coalition companion page
Overview
This October 2026 report presents nine practitioner-led action-research projects conducted in English schools during the 2025/26 school year. Its central message is that AI outcomes depend less on the tool than on the surrounding pedagogical routines, staff capability, leadership, school culture, data practices, and iterative implementation.
Six Findings
- AI worked as part of structured pedagogical routines, not as a standalone intervention.
- Independence and agency could move in different directions.
- Pupils often valued AI for bounded tasks but trusted teachers more for contextual and relational feedback.
- AI-marking reliability depended on carefully designed instructions and processes; long-form marking remained unsettled.
- Faster feedback moved assessment closer to teaching while raising questions about validity and what assessment measures.
- Delegating work to AI produced time savings alongside losses in teacher contact with pupil work and pupil attention to process.
Why It Matters
The report is useful as implementation evidence rather than a causal evaluation. Its case studies include younger pupils, SEND learners, disadvantaged communities, AI-assisted marking, reading support, early-language planning, and bespoke school tools. It argues for purpose-driven adoption, adequate training, local action research, and explicit attention to what automation displaces.
The repository bundle includes the full PDF, source metadata, and an interpretive synthesis alongside the archived source pages.