Introduction
MenTRnet is an international online network for mentors of teacher-research, with a particular focus on Exploratory Action Research (EAR). It is therefore a good forum in which to find out, in a preliminary fashion, how experienced and novice teacher-research mentors feel about the (potential) role of AI in supporting their work.
On 30 November 2024, Şirin Soyöz and I were invited to lead a discussion on this topic at a MenTRnet monthly meeting. We found that quite a few mentors are already experimenting with uses of GenAI and have a generally positive, while suitably critical, view. Others have not tried it yet but are interested in finding out more. Reflecting on these responses, we’ve decided to suggest setting up a learning community within MenTRnet for those interested in exploring AI use in mentoring teacher-research further, and we hope to get that going in February. For now, though, here’s our summary of discussion at the November meeting.
First, we briefly shared our own backgrounds in mentoring teacher-research and explained how we’d started discussing the use of AI in mentoring, together and then with the Noticing team, earlier in the year. We next asked participants to join breakout rooms to discuss the following questions;
- Have you or any teachers you’re mentoring used AI as part of research?
If yes, how? - Do you think AI could be useful to assist with (teacher-) research and/or mentoring of teacher-research? If yes, how?
Participants noted down responses in a shared Google document, and, following break-out room discussions, we asked a representative of each group to report back in plenary. Based on this data, here’s a summary of the main points made by participants:
Existing AI uses by mentors and mentees
GenAI (usually, ChatGPT, though Microsoft Co-Pilot was also mentioned) is already being used by some mentors or mentees for the following purposes (ordered here according to different stages of a typical EAR process):
– designing tools like surveys and questionnaires under mentor guidance
– entering data to generate thematic analysis
– assisting with literature search by processing and synthesizing large amounts of information.
– generating ideas for new strategies or techniques for the ‘action’ part of ‘action research
– generating teaching materials, stories, images (using ideogram.ai), questions and lesson plans for teaching when a new action had been decided upon
– writing up research findings
Another use, highlighted during the session by Sirin herself, was providing instructions (in the form of copiable text) which mentees could use as prompts for GenAI when direct mentoring is unavailable. She will be sharing some of these prompts in forthcoming blog-posts here which describe her recent mentoring experiences with AI.
Considerations when using AI
The mentors who shared the above uses tended to emphasize the importance of these uses being under mentor guidance, and of mentor and mentee both remaining sceptical or ‘critical’ about the work done by GenAI. For example, when using AI to generate themes from raw data, the AI-generated results should be reviewed by the mentor to ensure accuracy and alignment with contextual and research needs. One participant mentioned encouraging mentees to critically compare their own findings with AI-generated outputs, ensuring human oversight in interpretation of findings. In getting AI help in writing-up, too, the importance of authors retaining their unique linguistic styles was emphasized. Mentors also highlighted the importance of teachers adapting AI-generated strategies and materials (‘actions’ in action research) to fit the unique needs of pedagogical or research contexts. So, all in all, mentors stressed that it’s important to think of AI as a complementary tool rather than as a replacement for human creativity, insight, or expertise: “AI tools can process information and provide suggestions, but human oversight is very important to interpret these suggestions,” as one group wrote in their Google doc. Participants also highlighted needs for openness about AI use and compliance with ethical guidelines, as well as awareness about the possible (racial, gender, etc.) biases in AI output. Finally, a need for training in AI use both for mentors and for mentees was also mentioned, given that, as one participant put it, “the quality of output [from GenAI] is determined by the quality of the input”.
Possible future uses of AI suggested by mentors
In participants’ own words (in the shared Google document), here are some ideas for possible future uses of GenAI to support mentoring of teacher-research, again ordered according to EAR stages from beginning to end of a project:
“Could be useful to a mentor for finding out basic context / cultural background of mentees (cross-culturally)”
“How possible would it be to have a dialogue with ChatGPT about the issue the teacher is planning to explore?”
“AI can be used to reformulate/refine research questions to make them sound more comprehensive, clear, and concise.”
“Using AI … as a mentor-like resource for validating research methodology or exploring ideas for alternative research strategies.”
“Have a dialogue when looking for possible actions – ChatGPT can be used as a resource for pitching some ideas about possible improvements.”
And, it could be used as a mentor for mentors:
“[When] mentoring and we face a challenge, we could ask GenAI – compare our original thoughts with what we find.”
The overall feeling seemed to be that AI could – indeed, already can – be useful to assist with teacher-research and teacher-research mentoring but that knowing how to use it effectively is key, including providing it with “criteria and ‘right prompts’ and as much background information as possible” (one participant, in Google doc). Indeed, training in the use of AI in teacher-research (mentoring) could contribute to AI literacy and digital critical literacy development in general. As was also mentioned, however, mentors need to learn not only how to use AI themselves but how to introduce mentees to its effective use – so additions will be needed to the teacher-research mentoring advice already contained in publications like Mentoring Teachers to Research Their Classrooms: A Practical Handbook.
AI’s potential for increasing the number of teachers who can be mentored through the process of EAR was highlighted by some participants, and that is, indeed, one of the two major reasons why Sirin and I think it’s worthwhile to continue exploring in this area – the other being the value of improving individual mentors’ and teacher-researchers’ independent skills through work alongside AI. We’re looking forward to further collaboration with MenTRnet as a good way to take forward this thinking.
Prof. Richard Smith
Image Credit: ideogram.ai
