EEF/NFER Randomised Trial: ChatGPT Cuts Science Teachers' Lesson-Prep Time by 31%
United Kingdom
A school-randomised controlled trial of 259 Year 7–8 science teachers in 68 English schools found ChatGPT use cut weekly lesson-prep …
United Kingdom · Bristol · See the United Kingdom profile · See the Bristol profile
Top 40% 53/100 · Ask Evidence Copilot about this practice
Jisc piloted TeacherMatic with 8 further-education colleges (c.400 staff) and 7 universities (c.350 staff); 179 FE respondents reported saving 2h10 a week on average, while the HE pilot found 51% felt it saved time. Self-reported, with survivorship bias acknowledged.
Jisc, the UK's digital agency for tertiary education, ran two member pilots of TeacherMatic, a generative-AI platform with 50+ generators for lesson plans, quizzes, schemes of work and feedback.
Further-education pilot: eight colleges (including Basingstoke College of Technology, Milton Keynes College, NPTC Group and West College Scotland), chosen from over 60 applicants, each received 50 licences for 12 months, giving around 400 participants (May 2023 to February 2024). A mid-pilot survey drew 102 responses (87% reported saving time; 93% would continue). The final survey drew 179 responses: 79% said it saved time, with an average of 2 hours 10 minutes per week; 85% rated content accuracy very or somewhat accurate; 70% rated content quality good or very good.
Higher-education pilot: seven universities, around 350 participants (January to July 2024). 51% said it saved time creating teaching resources, with an average of about two hours a week; 63% would recommend it for HE; 70% rated accuracy somewhat or very accurate. The feedback generator was singled out as useful.
Honest limits: results are self-reported surveys with response rates well under half of participants, and Jisc itself flags possible survivorship bias. Participants reported repetitive activities, content too basic for detailed subjects, and presentations with irrelevant images and without screen-reader-friendly formatting. Pilot leads stressed that users must check generated content for suitability and accessibility, and some skilled users found general chatbots gave richer output. No learner-outcome data was collected.
Read the full analysis: https://teachermatic.com/wp-content/uploads/2024/05/Jisc-AI-TeacherMatic-FE-pilot-report.pdf
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
United Kingdom
A school-randomised controlled trial of 259 Year 7–8 science teachers in 68 English schools found ChatGPT use cut weekly lesson-prep …
United States of America
A teacher-facing generative-AI platform adopted by 6M+ educators (more than all US K-12 teachers) with 80+ tools that save ~10 …
India
Microsoft Research India and Sikshana Foundation built an AI copilot that drafts curriculum-aligned lesson plans in Kannada/English for Karnataka government-school …
Australia
A secure, department-built generative-AI assistant helps NSW's ~2,200 public schools draft lesson materials and differentiated resources; scaled statewide by 2025, …
Open full copilot Grounded in cited practices — always check the sources.