Task- and Training-Orientated GenAI: Defining two Categories of GenAI use in ITT and Teacher Education

The DfE report ‘GenAI in Education’ (2024) positively summarises the potential of GenAI in Education as these three aspects:

●        Reduce Teacher’s workload and therefore increase free time to focus on ‘excellent teaching’

●        Tailor educational materials

●        Support students with SEND.

Additionally, it acknowledges the ‘risks and challenges’ that will need to be addressed as the sector advances. These risks and challenges are arguably where most exposure and effort currently lie within institutions, as worry about academic integrity and appropriate usage on programmes are rapidly infiltrating the latest cohorts. This report moves away from these risks and challenges to look towards to the potentials of GenAI for Teacher Education, as identified above, and moves to deconstruct these themes into tangible products, tools and approaches that can be used by Teachers and that align with the ITTECF and Teachers’ Standards (DfE, 2012). In essence, to work for and with Teachers and their practice.

It could be postulated that the development of Aila, the Oak Academy’s GenAI Assistant, has the aim of moving towards this stage of ‘partial automation’. It has recently been released for general use as a consequence of the ‘GenAI in Education’ report (DfE, 2024). The tool guides the user through the lesson planning process and produces tangible lesson materials such as quizzes, slides and exit tickets, in addition to the lesson plan. Under the hood, it boasts of having a ‘9000 word prompt’ to improve accuracy of material and specifics of its operation are hard to find. When user-tested and asked (for the purpose of this report) to create a lesson for year 6 on Friction it produced slides with little detail and no pictures. The lesson plan it produced does not meet the Teachers’ Standards and would not demonstrate achieving many of the ITTECF (DfE, 2024) criteria. The areas in which it is deficient are also the areas that scenario-based learning for Teachers could address, as will now be explored.

When the stages to automation are integrated with the three core potential benefits of GenAI as described in the DfE GenAI report (2024), it could be suggested that GenAI in Education can split into two categories:

●        Task-orientated GenAI: Tangible outcomes are produced or processes undertaken, for example lesson plans, lesson materials, report generation, marking of essays etc. Reducing Teacher workload and tailoring educational materials.

●        Training-orientated GenAI: Scenarios are defined (e.g. a child with dyslexia who prefers quiet) alongside an action (create a speech on the Egyptians) for GenAI to appraise (how could this activity be adapted for this scenario). Teacher (or Trainee) reflects and learns by considering the suggested intervention or adaptation. Supporting students with SEND.

To highlight this distinction, consider the potential to ‘support students with SEND’. Aila does not ask or provide an iterative interface to consider how tasks can be adapted for individual needs. This does not fulfil the ITTECF’s (DfE, 2024) ‘Adaptive Teaching’ requirement as it is prescriptive and not responsive. However, training-orientated GenAI could fulfil this criteria because it is used by the Teacher for their professional development.

A teacher can ‘set’ a scenario: i.e. a pupil who identifies with ASD and is noise-sensitive (a barrier to learning, as in ITTECF 5.7 (DfE, 2024)). The teacher can ask GenAI how they could adjust their intended activity for the needs of the student. The prompt could look something like this:

“I am teaching year 9 gravity with the objective of contrasting weight and mass. One of my pupils identifies with ASD and is very noise sensitive. Considering the noise sensitivity, would getting pupils to weigh different objects be a suitable activity?”

The response (try putting it in your own GenAI!) results in multiple pedagogical considerations and learning points, many of which could be integrated into the targets of teacher practice.

This iterative approach could be used to ‘train’ the Teacher through trial and error, without having to expose real pupils to avoidable mistakes in practice. This therefore improves Teacher confidence and promotes a stronger teacher identity. See Hedlund (2024) for further description of how the professional development of a teacher involves iterative child-led continual adaptation and development of practice, related to a wider theoretical context. Essentially, that this process could be repeated for many of the ITTECF criteria.

References

Department for Education. (2023). Generative AI in education: Educator and expert views. London: Department for Education. Available at: https://assets.publishing.service.gov.uk/media/65b8cd41b5cb6e000d8bb74e/DfE_GenAI_in_education_-_Educator_and_expert_views_report.pdf  (Accessed: 7 November 2024).

Department for Education. (2024). Initial Teacher Training and Early Career Framework. London: Department for Education. Available at: https://assets.publishing.service.gov.uk/media/661d24ac08c3be25cfbd3e61/Initial_Teacher_Training_and_Early_Career_Framework.pdf  (Accessed: 7 November 2024).

Department for Education. (2011). Teachers’ Standards. London: Department for Education. Available at: https://www.gov.uk/government/publications/teachers-standards  (Accessed: 7 November 2024).

Hedlund, V. (2024). GenAI for Rosenshine's principles of instruction. Retrieved from https://www.teachergenaitoolkit.co.uk/blog/genai-for-rosenshines-principles-of-instruction. Accessed: 8 November 2024.

Staneff, T. (2024). ‘I, Teacher: The Five Stages to Unleashing Robot Educators’, Schools Week. Available at: https://schoolsweek.co.uk/i-teacher-the-five-stages-to-unleashing-robot-educators/  (Accessed: 7 November 2024)