AGI and Education
Personalised tutors, automated assessment, and curriculum redesign — the realistic picture of how AGI may transform learning at every level.
Executive summary
Education is among the domains where AGI may have the largest positive effect: high-quality, patient, personalised tutoring at near-zero marginal cost. The same capability also raises serious questions about assessment, attention, equity, and what schools are actually for.
What changes
Patient one-to-one tutoring has been the gold standard of pedagogy since Bloom (1984). It is now possible to deliver something close to it at scale. Frontier systems already outperform unaided humans on tutoring trials in mathematics and writing.
Plausible benefits
- Universal access to expert-quality tutoring.
- Faster mastery in foundational subjects.
- Better support for neurodiverse learners.
- Lower workload for teachers on routine grading.
Credible risks
- Hollowing out of human relationships in learning.
- Assessment integrity in a world where students can co-write with capable AI.
- Widening gaps if access to high-quality systems is unequal.
- Atrophy of skills students offload to tools too early.
What helps
Curriculum redesign around things AI cannot do for students (collaboration, judgement, ethics, embodied skills); assessment shifted toward in-person and process-based formats; explicit AI-literacy as a core subject.
Key takeaways
- 01Personalised tutoring at scale is now technically feasible.
- 02Assessment models built for closed-book exams need redesign.
- 03Teacher roles shift toward mentorship and coaching.
- 04Equity of access is the largest open question.
Frequently asked questions
Should children use AI tutors?
Used well, yes — for explanation, practice, and feedback. Used to bypass thinking, no.
How should schools handle AI in homework?
Most educators are shifting toward in-class assessment, oral defence, and process portfolios rather than banning tools.