Chapter 4. Planning with a Tireless Assistant
In the summer term of 2024, science teachers in 68 English secondary schools took part in an unusual experiment. Researchers split the schools at random. Teachers in one half planned their Year 7 and 8 science lessons (students aged 11 to 13) with ChatGPT and an online guide, while teachers in the other half planned without any generative AI. Each week, both groups recorded how long their planning took (Roy et al., 2024).
One experienced teacher, whom the report calls Teacher B, had a low-attaining class that the school’s ready-made lessons did not suit. She used ChatGPT for about nine in ten of her Year 7 and 8 lessons. Her prompts described the class and what the students had already learned, and she followed up with requests to refine what came back. She also asked colleagues to share prompts that worked, because ‘once you’ve got the prompt right, you will save time’. ChatGPT saved her some time, but she valued something else more. ‘Students are getting a much better deal, having spent the same amount of time’ (Roy et al., 2024, p. 75).
What the chapter covers
- What the trials measured
- Hours or minutes
- Where the time goes
- Teacher-design burden
- The plan is not the lesson
- Differentiation, SEND and paperwork
- Teacher-facing first
- Who gets the saving
- What to hand over, and what to keep
- In practice
- When resources are thin
- Questions for reflection
- Framework connections
Questions for reflection
Take these questions to your next department or phase meeting, or into a quiet moment after school.
- In your own school, which planning task would you refuse to hand to a machine, even one that did it well?
- When did a generated resource last take longer to fix than it saved?
- What would go on a context card for your most demanding class, and what must stay off it?
- If your team saved an hour a week, who would decide where that hour goes?
- How will the newest teachers in your school learn to plan if the first draft always comes from a machine?