We launched the AI-in-Mentoring Special Interest Group (SIG) within MenTRnet in February 2025 to explore AI’s role in supporting teacher-research and mentoring for Exploratory Action Research (EAR).
At our 22nd February meeting, research mentors discussed AI’s potential in topic selection and research question development, based on some pre-meeting experimentation with various AI tools. While AI tools like ChatGPT, DeepSeek, Claude, and Gemini helped generate research ideas, participants had found the overwhelming number of options challenging. Ethical concerns, including AI hallucinations and bias, were also highlighted. AI often suggested action-oriented rather than exploratory questions, reinforcing the need for critical evaluation to maintain the exploratory nature of EAR.
Our final key takeaways from the meeting and reflection forms were as follows:
AI’s Role in Supporting Teachers and Mentors
-
- AI can act as a personal mentor, especially for large groups of teachers, but requires proper training and oversight.
-
- Mentorship and training are essential, especially for new teacher-researchers who could feel overwhelmed by AI-generated information.
-
- A customized GPT, designed to use only information provided by the user, reduced irrelevant responses and showed promise for guiding teachers in Exploratory Action Research (EAR).
Importance of AI Literacy and Prompting
-
- Teachers need AI literacy skills to engage with AI tools effectively.
-
- The effectiveness of AI tools depends on clear, structured prompts. Follow-up prompts are necessary to refine and narrow down research questions. Teachers need to provide clear direction to get relevant responses.
-
- Teachers might focus too much on crafting prompts instead of addressing classroom issues, highlighting the need for balance.
Limitations and Ethical Considerations
-
- AI tools can help with brainstorming of topics and draft questions but may not enable sufficient depth or contextual relevance
-
- AI-generated research questions can be too general or action-oriented, requiring human refinement for exploratory research.
- AI language needs to be simplified for accessibility.
-
- We need to be mindful of ethical concerns, verifying AI outputs for accuracy and fairness.
-
- We need to balance AI-generated outputs with human judgment to maintain critical thinking.
Moving forward, we will explore AI tools for selecting and designing data collection methods and continue working with partners to refine AI’s role in research mentoring.
Some of the questions that we will explore further:
-
- How can new teacher-researchers use AI as their personal mentor without feeling overwhelmed by responses when they have limited prior knowledge?
-
- What strategies can mentors use to guide teacher-researchers in crafting effective prompts?
-
- How do we ensure AI-generated responses remain relevant and aligned with the exploratory nature of EAR, rather than leading to action-oriented solutions too soon?
-
- How can we encourage mentees to critically evaluate AI-generated outputs rather than just passively accepting them?
-
- How do different AI tools (e.g., ChatGPT, Gemini, DeepSeek, AI Pro) compare in their ability to support the EAR process?
-
- What challenges do multilingual teacher-researchers face when using AI for EAR, and how can we address linguistic issues?
This blog post summarises key insights from our AI-in-Mentoring SIG meeting and reflections shared by participants. AI tools, including ChatGPT and DeepSeek, were used to synthesise and organise key themes from reflective written material and discussions, ensuring a comprehensive overview of our findings.
Şirin Soyöz Yilmaz
Image credit: https://ideogram.ai/
