An AI-Powered Lesson Planning Assistant with Contextual Curriculum Alignment

• In Event: Curating Educational Content

Proposal

Record numbers of UK teachers are quitting the profession, with workload acting as the main contributor. Alongside marking and assessment, teachers spend significant amounts of time planning. Our AI-powered lesson planning assistant that allows teachers to create personalized lesson plans and resources in minutes, was launched in September 2024. Initial teacher testing shows that it saved teachers around 3.5 hours a week.

Piloting of the first experimental AI tools began in October 2023. The approach exemplifies the UK government’s approach to AI in education, drawing on curated high-quality content to support lesson planning without compromising quality and safety. The lesson planning tool is designed to support, not replace, teachers’ expertise nor remove their agency and maintains the integrity of the lesson planning process. Teachers are given agency to use their expertise to make changes such as adapting literacy levels or activities, both during the process and after download. It creates resources that are ready for teachers to use, such as lesson plans, teacher slides, pupils’ quizzes, worksheets and practice tasks.

Evidence shows that teachers remain concerned about the accuracy, bias and safety of AI. Until now, generative AI tools have used large datasets from the internet for their content, sometimes drawing on materials that are inaccurate and unsuitable to use in the classroom. With our tool, we start to tackle this by pulling content from our existing curriculum resources, each created and checked by teachers to improve accuracy. Another issue with AI tools is their lack of understanding of context and local best practice. The tool has been engineered to understand the English national curriculum and tailors resources accordingly.

Whilst already tested by thousands of teachers, the public beta release of the tool facilitates the gathering of much greater feedback to feed into the development cycle. Continual evaluation will check the quality and performance of the resources generated. Furthermore, the tool can support the wider market to innovate and create safe AI tools. The code is available on an MIT license, and the underlying content and resources are Open Education Resources (OER) available on an Open Government Licence (OGL), permitting any developer to use the underlying material. This includes an Application Programming Interface (API) to allow companies to build off, adapt or integrate content used in the tool.

Our presentation will showcase an example of the tool in use and discuss further details of its development and key learnings from wider-scale usage.

Author

• John Roberts, Oak National Academy