# A new compact for human learning

FRONTIER THOUGHT · 2026-11-22



Editorial source review and next moderator brief



## What this round established

Test one lesson agreement: permitted help during practice, an unfamiliar unaided task, a later transfer check, an accessible alternative and a teacher-time budget. The evidence makes the teaching arrangement, rather than access to a generic chatbot, the intervention to compare.



## Comparative analysis

The Day 2 comparison did not produce a verdict for or against every tutor. It identified teaching design as the intervention and independent performance as a necessary outcome. The next step is a learning agreement that a school can explain to parents and pupils. It should name the capability being developed, the assistance permitted during practice and the evidence required afterwards. The cited experiments concern particular settings and short learning windows. Neither supplies a national threshold for Greek schools. Our proposed agreement therefore adds a delayed, unfamiliar task and a check on who benefits or is left behind. Those are questions to investigate, not outcomes already established. The agreement also needs a practical teacher budget. If every lesson requires extensive prompt preparation and checking, expansion without paid time could erode the support that made the intervention useful. A school should compare the whole teaching arrangement with an appropriate alternative, such as teacher-led practice, rather than compare an elaborately designed tutor with no support. The moderator’s next question is whether the school can defend the same learning objective for a pupil who needs another route. This moves the discussion from excitement about a tool to a testable educational commitment.



## What remains open

Delayed transfer, effects for different pupil needs and sustainable staffing remain open. No new Greek pilot result or national success threshold is available here.



## The next question

Can a Greek school promise stronger independent thinking for every pupil, and resource the teaching needed to make that promise credible?



## Published answers already available



### Sal Khan

How AI could save (not destroy) education · 2023-04

https://www.ted.com/talks/sal_khan_how_ai_could_save_not_destroy_education?view=transcript

Khan proposes personal tutoring and teacher assistance designed to prompt reasoning and give feedback rather than simply supply answers. His demonstrations are a design proposal, not proof of learning gains in Greek schools.



### Michalis Bletsas

Interview on AI and learning on ERTnews Syndeseis · 2026-08-26

https://cyber.gov.gr/i-techniti-noimosyni-mporei-na-enischysei-ti-mathisi-alla-yparchei-kindynos-gnosiakis-atrofias-synenteyxi-toy-dioikiti-tis-eak-m-mpletsa-stin-ekpompi-syndeseis-ertnews/

Bletsas warns that outsourcing a student’s thinking can weaken learning. He also sees value in personalised exercises and in teaching students to detect model mistakes.



### Bastani and colleagues

Generative AI without guardrails can harm learning: Evidence from high school mathematics · 2025-06-25

https://doi.org/10.1073/pnas.2422633122

A Turkish school experiment found that unrestricted assistance could improve practice while weakening subsequent unaided performance. A teacher-informed tutor largely mitigated that harm. The study does not establish long-term effects in Greek schools.



### Kestin and colleagues

AI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting · 2025-06-03

https://www.nature.com/articles/s41598-025-97652-6

A purpose-built tutor improved immediate learning outcomes in a Harvard undergraduate physics trial. The intervention combined teaching design and AI; it does not establish that a generic chatbot replaces a teacher or produces lasting gains.



## The people and evidence for the next exchange



### School leader and teacher

Which capability, alternative lesson and preparation time will you guarantee before expansion?

Draft learning agreement, timetable, staff-hour budget and review date.



### Independent education evaluator

Does the comparison isolate the teaching arrangement fairly, and which delayed task would test transfer?

Baseline and comparison design, common rubric, missing-participant record and data-minimisation plan.



### Student-support representative, pupil and parent

Is the goal clear, can help be challenged and can a pupil learn through the alternative without a penalty?

Plain-language explanation, an example of independent work and a documented support route.



## Moderator handoff

Use a teacher-led mini-roundtable to review the Day 2 lesson protocol. Ask for one assisted example and one independent explanation, shared only through the school’s approved process. Record which criterion changes and why. Put the revised agreement into the next brief beside the original; the organiser chooses the follow-up date.



Record the contribution, what it changes, what remains uncertain and the next question. Preserve the earlier brief. Published source material is distinct from any new participant response.
