During the summer of 2026, I had the opportunity to undertake an EDI research internship within the School of Psychology, investigating whether generative AI can be used to promote feedback literacy in higher education, with a particular focus on metacognition and on supporting students from underrepresented backgrounds, including neurodivergent students and those for whom English is an additional language.
I came to this project as someone who has sat on the receiving end of feedback that left me none the wiser about what to actually do next. That experience, which is far from unique, made the research feel personal. Feedback is one of the most powerful tools available in higher education, yet so many students, particularly those who already face additional barriers, struggle to translate what they receive into meaningful action. I wanted to find out whether AI could truly support students in becoming more reflective, self-directed learners, rather than simply doing their thinking for them.
My starting point was building a solid evidence base. Working alongside Dr Heather Branigan and Dr Kevin Allan, I read and synthesised literature across three connected areas: research on generative AI and metacognition, feedback literacy theory, and the experiences of underrepresented students in higher education. The intersection of all three turned out to be surprisingly under-researched, particularly when it came to understanding how race, socioeconomic background, or neurodivergence specifically shapes the way students engage with feedback. That meant a lot of careful reasoning from adjacent literature rather than finding a pre-existing answer. Although challenging at times, it was also extremely interesting.
From that foundation, I produced a framework proposing a five-stage sequence for an AI agent designed to promote feedback literacy. The distinction that mattered most was between an agent that generates polished responses for students and one that prompts students to do the cognitive work themselves. The goal was always the latter. A key finding from this stage was that the agent performed exceptionally well at eliciting metacognitive reflection - getting students to think about how they were learning, not just what they got wrong. Seeing the framework I had built actually shape the behaviour of a working tool was one of the most satisfying moments of my academic life so far. The project has since moved into refining the prototype further: making the learning process more conscious for students, supporting the transition from emotional reaction to constructive action, and ensuring the tool scaffolds thinking rather than replacing it.
Beyond the research itself, this internship pushed my skills in ways I did not entirely anticipate. Pulling together disparate literatures into something coherent and usable is genuinely hard, and doing it at this scale was a first for me - an experience that will absolutely inform how I approach my honours project. Writing progress updates and a summary of my research also sharpened my ability to communicate complex, in-progress work clearly to an audience with limited time, a skill I will carry beyond academia.
The biggest challenge was sitting with uncertainty. I came in with no prior knowledge of metacognition in the context of education or of feedback literacy, and the topic initially felt quite daunting. But that turned out to be one of the most valuable parts of the experience. Going in without fixed assumptions meant I remained open to what the evidence was actually saying, including when it complicated rather than confirmed what I expected.
If you are considering applying for an EDI internship, whether or not you know anything about the topic, I would highly encourage you to go for it. Interest and curiosity, and the persistence to keep going when things do not immediately click will take you far. And do not hesitate to ask your supervisors for help as that is exactly what they are there for.
I took away sharper research skills, a genuine new interest in education as a field, and the knowledge that the work I contributed will go on to support students who could really benefit from it. That made it more than worth it!
Megan is currently entering her 4th year of the BSc Neuroscience with Psychology programme at the 51cg.
