From the SLED-wide archive edition of August 29, 2026
Controlled school trial cuts lesson-planning time without a detected quality loss
Education Endowment Foundation and NFER · K–12 education · England, United Kingdom
- Publisher
- ChatGPT in lesson preparation – Teacher Choices trial
- Original publication
- August 2026
- Source retrieved
- Not recorded in the historical archive
What happened
An independently evaluated Teacher Choices trial compared ChatGPT-assisted and unassisted lesson and resource preparation among 259 teachers in 68 state-funded secondary schools.
Why it matters
This is unusually strong evidence for a bounded education workforce use case: reduce teacher preparation time while separately checking resource quality rather than assuming faster output is better instruction.
Evidence and measured results
ChatGPT-group teachers spent 56.2 minutes per week on relevant planning versus 81.5 minutes in the comparison group—a 25.3-minute, or 31%, reduction. A blinded expert panel did not detect lower resource quality, and the time result received a high security rating.
Limitations and uncertainty
The trial covered Year 7 and 8 science planning, not student use or learning outcomes; schools slightly overrepresented London, the South East, and highly rated providers, and frequency of tool use declined during the trial.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source as summarized in the preserved archive. Enriched 2026-09-05; this does not change the original publication date. Labels below come from the analysis itself.
Sales
Role takeaway
- Customer problem
- teachers need to manage lesson-preparation workload without sacrificing the quality of instructional resources.
- Stakeholders
- district curriculum and science leaders, teacher-workforce teams, school leaders, IT, and privacy.
- Discovery
- which preparation tasks match Year 7–8 science planning; how is time measured; and who can review resource quality without knowing how it was produced?
- Value hypothesis
- supported teacher-facing preparation may reduce time while preserving quality under local evaluation.
- Potential engagement
- a limited curriculum-aligned workload pilot using the trial's separation of time and quality measures.
- Unsupported claims
- the reported 31% reduction and undetected quality loss do not establish improved attainment, benefit in all subjects, equivalent U.S. results, or safety of learner-facing use.
Pre-sales engineering
Role takeaway
- Fit
- start with teacher-facing planning and resource preparation, keeping generated materials in the normal curriculum review path.
- Architecture and integration
- provide an approved environment, curriculum context, reusable prompt examples, and a simple way to retain drafts for review; complex classroom integration is not established as necessary by this record.
- Prerequisites
- defined tasks, a quality rubric, teacher guidance, and baseline preparation-time measures.
- Constraints
- the evidence covers specific science year groups, and use declined during the trial.
- Security
- keep student personal information and protected records out of general-purpose prompts and test the training examples for that boundary. Proposed proof: compare preparation time and blinded resource-quality ratings locally, reporting workload and student outcomes separately.
Delivery
Role takeaway
- Work
- agree the preparation task, give teachers the guide and prompt examples, allow a learning period, and collect time and independently reviewed resource samples.
- Dependencies
- curriculum reviewers, teacher participation, privacy-approved input examples, and scheduled learning time.
- Ownership
- curriculum leads accept resource quality; a workforce/benefits owner evaluates workload; teachers remain accountable for classroom materials; IT/privacy teams govern the environment.
- Skills and adoption
- practice adapting and checking drafts rather than accepting them unchanged.
- Governance checkpoints
- task/rubric agreement, pilot review, and evidence review before new subjects or student use.
- Proposed acceptance
- compare preparation time with baseline while meeting the agreed quality rubric and input rules. Risks include declining use and inferring student learning improvement from faster teacher preparation.
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Start with teacher-facing, low-risk preparation workflows; retain the guide and prompt examples as part of implementation; and keep generated resources in the normal human review and curriculum-management path.
Governance
Who approves, reviews and stays accountable for outcomes?
Define the intended task and quality rubric before deployment, allow a learning period, and measure workload and instructional quality separately from student attainment.
Security and privacy
What data, permissions and controls need testing?
Keep student personal information and protected records out of general-purpose prompts, and include approved data-handling examples in teacher training.
The preserved archive analysis covered architecture, governance and security. Not assessed for this record: accessibility and workforce, procurement, operating model.
Publication history
- 2026-08-29SLED-wide archive · Issue 0214 resources
Stable resource ID: eef-lesson-planning-trial