AI Week: The Adaptability of a Values-Based Reflective Practice Model with Generative AI: A Case Study

In this blog, Danny Clegg, Associate Professor of Digital Pedagogy & Inclusive Practice at BPP University, explores how a values-based reflective practice model in healthcare education was adapted for use with Generative AI, making reflection more inclusive, emotionally safe, and empowering for diverse learners.

About the Author:

Danny Clegg is a neurodivergent Associate Professor of Digital Pedagogy and Inclusive Practice, and Associate Dean of Academic Development for Healthcare and Nursing at BPP University. He is also a Director of the National Association of Disability Practitioners. His work is driven by a commitment to Universal Design for Learning (UDL), radical inclusion, and challenging power-deficit models in both technology and education. Through values-based innovation, he ensures those most at risk of exclusion are empowered from the start – not added in as an afterthought.

That’s how I describe the moment when a values-based reflective practice model I had been developing since 2021 – was introduced to Generative AI (GenAI) in 2024. As a neurodivergent educator and developer of inclusive pedagogies, I’d long worked on reflection approaches in academic settings that welcomed and valued difference, rather than tolerated or devalued it. What I hadn’t predicted was how easily and intuitively those same principles would transfer into an AI-enabled framework.

This case study aims to share some of the experiences and the motivation behind exploring how a modernised academic and professional practice developed in healthcare programmes, married up so cohesively with GenAI.

---

Why Values-based Reflective Practice? And Why Now?

Reflection is a core part of learning in healthcare and professional education. But traditional models often privilege academic language, structured thinking, and dominant cultural norms. For many learners – particularly disabled students, neurodivergent learners, or students for whom English is an additional language – reflective practice can feel like a gatekeeping task, with the core essence and value in the practice sinking in an ocean of outdated marking criteria.

Over four years, I developed a values-based reflective practice framework designed to:

• Invite learners to reflect on what matters to them
• Reduce barriers caused by rigid academic models
• Honour emotional responses without penalising authenticity
• Align with the reflective practice values and domains learners will engage with post-qualification (e.g. Professional Registration CPD cycles)

What this modernised model of reflective practice aims to achieve is not to reinvent reflective practice, but to reimagine it as a patchwork development of skills, moments, and values – making it real, reachable, and relatable.

---

From Classroom to Chatbot: Reflective Practice Meets GenAI

When Generative Pre-trained Transformer (GPT) tools became available for potential use and development in educational settings, I explored how the reflective practice model could be tested within a GenAI environment. Rather than crafting a new model for AI, I embedded the model into the interaction design of a bespoke GPT I had built to mimic the core values held behind an EDI-focused lens, as a live classroom tool to simulate reflective conversations with learners – co-developed and trained with and for learners.

Because the model is inherently values-centred it translated well to a GPT where the learner was not just typing prompts or frantically checking they had correctly placed their content under the expected reflective practice model subheadings (something routinely found in marker feedback correlating with lower grade outcomes), but reflecting in a non-threatening space, through dialogue. Importantly, I integrated automated halts for emotional disclosure – meaning if a student expressed distress, the GenAI immediately paused the session and provided signposting prompts to support routes. This drew directly on the ethical safeguards I’d developed in focus group methodologies over many years prior to first exploring AI technologies.

---

Learning Through Emotion, Not Around It

One of the most powerful components of values-based reflection is its emotional honesty. But that also brings risk.

That’s why my reflective practice approach insists on pausing and acting when distress is detected, whether in human-led or AI dialogue. If reflection becomes emotional labour, there must be safeguards. AI can support that – but only if programmed with intentionality.

Too often, reflective practice is treated like a box-ticking task. One learner shared:

“At the start of this module I hated reflective practice with a passion… I hadn’t realised that each of the activities were actually components of reflective practice using the DEEP reflective practice approach, until the ‘big reveal’ in the final session! Very sneaky clever of you Mr Danny… but I’m glad to leave this module having enjoyed reflective practice for the first time!”

For some, it wasn’t just about enjoying the process — it was about seeing their values recognised:

“I didn’t just learn how to reflect, I learned how the things that are important to me were genuinely valued in the feedback… Best of all, I learned to listen to the experiences of others… using this GPT has helped to open my eyes to the diversity of other students on my course, and how even an AI can behave with cultural competence – so there is no reason why I can’t too!”

---

AI Isn’t a Replacement. It’s a Reflective Tool – With Guidance.

AI-enabled reflective practice is not without pitfalls:

• Learners may rely on AI to generate, rather than explore, ideas
• Emotional content could be mishandled by generic GPT systems
• The risk of “tech solutionism” – assuming AI can replace human support

That’s why it’s vital to build in safeguards, structured prompts, and human moderation. Learners must be taught to use GenAI critically, as this student noted:

“I was terrified of AI and how risky it could be… but when I realised that the dialogue I had been having with the AI then gave me the skills to critically analyse the content… I felt empowered to continue using it – but only use it responsibly.”

With the pace of the race, and so much of the focus on quality of GenAI outputs, It’s a discussion that needs urgent attention in higher education. As highlighted in a recent WONKHE blog:

“Higher education needs a plan in place for student pastoral use of AI”
– WONKHE, Sept 2024

This work responds to that exact gap: recognising that students are using GenAI for reflective and emotional learning – and ensuring there’s ethical, inclusive, values-based infrastructure to support them.

---

What’s Next?

The model of reflective practice has been integrated with the standard placement reflective practice parameters involving the NMC pillars, to create a reflective practice GPT, currently in the pilot testing phase, to be used by Nursing students as part of a social impact placement project, funded by the London Higher Education Group (LHEG), in partnership between BPP University and London South Bank University. (Keep an eye out for the published outcomes of this initiative in early 2026!)

---

Final Reflections

In a world where GenAI will increasingly intersect with learning, it’s not just what we build that matters – it’s who we centre as the user, and how we value their participation in the designing and development stages.

The reflective practice model was never about technology. It was about equity. And it was most certainly never about creating a product. When we start with values – respect, authenticity, safety – we don’t need to “adapt” for inclusion. We build it in from the beginning.

And when that meets AI? It doesn’t just scale. It empowers.

---

Citation of the DEEP model (pre-print abstract only):

Clegg, D. (2024). Embedding Values-Based Reflective Practice into GenAI Frameworks: The DEEP Reflection Model. Zenodo. https://doi.org/10.5281/zenodo.14421650

AI-Transparency statement: The author used a Generative Artificial Intelligence (GenAI) tool to assist in the preparation of this blog article, but reviewed and commits to every word of it – holding themselves accountable for the content, and ensuring the views expressed within are their own and not necessarily those shared by any individuals or organisations they are associated or affiliated with.