AI for Learning: Helping Foundation Year Students in the Physical Sciences use ChatGPT Effectively

Project members

Christopher Patrick, Rachel Quarrell, Nicole Miranda (MPLS Division).

Project summary

Using AI to help Astrophoria Foundation Year physical sciences students create personalized learning materials and integrate AI-based tasks into their studies to boost confidence and understanding. 

View final project report (PDF)

AI in Teaching and Learning at Oxford Knowledge Exchange Forum, 9 July 2025

Findings from projects supported by the AI Teaching and Learning Exploratory Fund in 2024–25 were presented at the AI in Teaching and Learning at Oxford Knowledge Exchange Forum at Saïd Business School on Wednesday, 9 July 2025.

Project team members each presented a lightning talk to all event participants, and hosted a series of small group discussions.

Follow the links below to view the lightning talk recording and presentation slides for this project.

Click image above to play lighting talk recording

View presentation slides (PDF)

Project case study

A lot of public-AI use amongst undergraduates is currently very passive, which can suppress the development of student skills and abilities.   

 The Astrophoria Foundation Year (AFY) AI project was devised to investigate the possibility of creating positive tools for active learning and revision which would aid students on the CEMS (physical sciences) course in improving their own academic skills, instead of using AI as a crutch to replace skills they were meant to be developing during the Foundation Year.  As part of this it was planned to brief and induct students on positive AI use, and to warn them about the dangers of passive reliance on AI.  

The AFY CEMS students and a group of interested Astrophoria science tutors were briefed by Kelly Webb-Davies of the OERC, and three possible options were then investigated.      

1. Use of the AI as a Socratic tutor, with which students could ‘converse’ about difficult science concepts, but encouraged to answer the question themselves rather than just be given the answer.  This mimics the style of active learning prevalent in Oxford undergraduate tutorials.  
2. Use of the AI to self-quiz and find holes in knowledge or conceptual understanding, broadening the amount of self-test material available to them from a new course.  
3. Use of the AI to generate a paragraph of text about a CEMS topic, within which the students had to find an unknown number of errors (grammatical, logical or factual) without the AI giving them answers until they were completely correct.  This is a form of active learning not often used at Oxford but it can be very powerful for clearing up misunderstandings and ensuring that students learn to read exam questions and textbooks carefully.  

By the end of the project GPTs had been developed for all three methods and tested by the students, course leads and tutors.  

All three methods worked well, though the AI persistently tried to provide full answers.  Students gave feedback which offered new insights into the undergraduate use of and opinion of AI.  Tutors and course leads learned a lot about how to discuss AI with students, and how to use it themselves.  

We would like to continue this approach with the next cohort if possible, and to broadcast more widely the need to ensure that pre-graduation students are using AI for active learning rather than passive use, which can prevent their skills improving.