AI Teaching Assistant (AITA) Tool (School of Engineering)

This project builds  on previous work in the School of Engineering which involved the development of an AI-Teaching Assistant (AI-TA) for an MSc module on “Motion Picture Engineering”, which was trained specifically on module content. This project will develop a structured workflow for rolling out AI-TA across multiple modules. It will also refine assessment and feedback tools within  the AITA framework with a long-term view to rolling it out across the School of Engineering.

• Roll out AITA to the Wave Energy and Motion Picture Engineering modules
• Develop strategies for using AITA to summarise student feedback to lecturers in a way that allows the lecturers to address learning issues continually as the module is delivered
• Develop verification methodology to check trustworthiness in teaching environment (assess AITA based on its performance on previous examinations)
• Develop “best practice” guidelines for using AITA as a TA
• Develop “best practice” guidelines for using AITA in student assessment

Phase 1 Oct – Nov 2025 GUI Ingest

Workflow Development: unified single point of entry to simplify this process into one step of upload followed by a series of automated steps which do not require supervision.

Phase 2 Nov2025 -Jan 2026

Assimilation and Ingest of Wave Energy module by Professor Basu

Phase 3 Jan – April 2026

Deployment Alpha-1 Version

Phase 4 Feb – April 2026

Prompt engineering to create a weekly summary of student interactions with AITA recommending to the lecturer material that should be revisited.

Phase 5 April – Sept 2026

LLM refinement testing: Deploy simple refinement strategy on the uploaded materials to improve accuracy of Chatbot responses.

Phase 6 June – Oct 2026

Guidelines for use of AI-TA as part of a Journal paper in Education

Phase 7 March – Dec 2026

Evaluation: Development of impact assessment tools and trustworthiness exercises.

Phase 8 Sept – Dec 2026

Development of assessment infrastructure: Prompt engineering for assessment exercises, and development of a protocol for use of AITA in open book in-class tests

Phase 9 Jan 2026 – March 2027

Launch refined workflow (Beta Version AITA)

Software system for running AITA and standardised GUI for interaction

This system will be deployable across School.

Workflow for using AITA in any course

 This is the operational manual for how the system is to be used by staff every day.

Guidelines for verifying trustworthiness of AI-TA

This gives us a set of standard “examinations” which we can use to verify that AI-TA responds with reasonably accurate answers for specialised subject matter.

Journal paper on Education with AI-TA

This allows us to disseminate our findings internationally and begin the journey to wider impact.

Best practice guidelines for use as student feedback tool and for use in assessment

Key deliverable for usability by academics

Analysis tools for evaluating impact of AI-TA and use by students

This is the suite of software tools for post-analysis and visualisation of impact

Student performance comparison by evaluating with and without AI-TA

Self-evident. This will convince academics and students about the utility of AI-TA

The non-technical outcomes will be developed in partnership with the Centre for Academic Practice.