AI could be used to analyse recordings of lessons delivered by new teachers, to give feedback on what they are doing well and what could improve, as part of a National Institute of Teaching (NIoT) project.
Raj Chande, senior research associate at the NIoT, said early career teachers’ (ECTs) lessons would be recorded and uploaded to “a secure tool that’s…converted into a transcript”.
“Those transcripts are going to be analysed by…relevant experts who are going to be able to say ‘that’s great questioning there’ or… ‘they shouldn’t have started that explanation with that much disruption going on’.”
The feedback given by human experts will then be used to train an artificial intelligence (AI) model, to “replicate the judgement of experts”.
“At scale, mentors and ECTS [could] be given semi-automatic feedback, all quality assured by our team development experts.”
Chande said that if the project is successful, teachers could be “given instantaneous, semi-automatic feedback without anybody actually having to come into the room necessarily.”
He said the scheme is not designed to “substitute…having a conversation. But we’re trying to reduce the constraints on time and trying to increase the breadth of lessons that are observed.”
But the project would allow the establishment of “a database of teacher practices” comprising “thousands of lessons”.
“We [will] know what’s being done and what’s not being done quite so well.”
The project was discussed at the Confederation of School Trusts (CST) annual conference in Birmingham today.
Chande told Schools Week work on the project will begin this academic year, with around 12 ECTs testing the recording tool to ensure it works.
Far higher numbers would then record their practice next academic year, to build up more than 1,000 lesson transcripts.
It is understood conversations are currently underway around the legal aspects of the programme.
Calum Davey, executive director of Research at the NIoT, said: “Observation and feedback are central to teacher development. In this project, we are investigating whether AI can be trained to identify classroom practices well enough to be useful for supporting teacher development.
“We’re committed to understanding what the benefits of AI tools are and equally committed to asking the difficult questions about bias, judgement, and the kind of role we want technology to have in our schools.”
Al is computer generated: computers are only able to work on data they are fed. If the input is rubbish and the agenda of those who feed the information in flawed then whatever AI generates is flawed rubbish that reinforces the bias and myopic thinking that it receives.
OFSTED and the UK education model is not for purpose. We have a regimented system that it based on a batch production factory model, current teacher training eliminates creativity and reinforces the narrow silo’d subject departmental system. This proposal is a misuse of AI, far better to use the power of AI to tailor learning to the needs of individual students.