
Dr Antony Hill is the Dean of Learning & Teaching for the College of Health, Science and Society at UWE Bristol. In this post, he considers whether artificial intelligence (AI) is helping to create a more inclusive higher education sector or introducing new barriers for learners, arguing that the answer depends less on the technology itself than on how universities choose to implement it.
The question is not if Artificial Intelligence (AI) belongs in higher education. The question is what kind of higher education we want AI to support.
Conversations about AI in higher education tend to become polarised very quickly. AI is often presented either as an all-encompassing solution to long-standing educational challenges or as a threat to academic integrity, quality and human connection. Of course, like most things in life, it’s not one or the other; it’s a bit of both (as explored in a previous article From disruption to deliberate practice: AI in Higher Education).
Part of the tension comes from what AI challenges.
Questions around authorship, and what this means when an ‘output’ is so easy to generate, sits at the heart of current debates. Assessment often becomes the first site of anxiety because it is where institutions attempt to preserve trust, standards and evidence of learning.
Framing AI as either entirely positive or entirely harmful risks missing a highly important, more nuanced, question: under what conditions can AI support inclusive education?
AI systems can reflect design choices, data histories and pedagogical values. Inclusive institutions may use AI to widen opportunity. Less reflective approaches risk reinforcing existing inequalities.
Where AI can support inclusive education
Used thoughtfully and intentionally, AI has the potential to remove barriers and improve participation. Examples already emerging across higher education include:
- text simplification and translation support
- study support and conversational tutoring
- automated captioning and transcription
- formative feedback and personalised scaffolding.
These approaches can particularly benefit students who are first in their family to attend university, students studying in an additional language, commuter students and those managing work or caring responsibilities.
Reducing time spent on routine administration, formatting and repetitive drafting may also allow us to invest more in mentorship, dialogue, responsiveness and relational teaching; aspects of learning that remain distinctly human.
Where AI risks becoming a barrier
Inclusive education, however, requires a focus on who is actually benefitting and who may be excluded.
Educators and students are already using AI but not equally well, critically or consistently. Differences in AI literacy, access to premium paid tools (creating unequal access) and institutional guidance can create hidden advantages and disadvantages. Banning AI does not remove these inequalities. In some cases, it simply makes them less visible.
Another concern is homogenisation. If everyone relies on the same generative systems, there is a risk of flattening creativity, reducing disciplinary voice and encouraging formulaic outputs. In a similar vein, AI systems are trained on existing data and social patterns, which means they can reproduce bias at scale.
Crucially however, the challenge is not really with AI itself. It is the limited support/clarity around it, and uncritical adoption.
Moving from restriction to responsibility
A shift from restriction to responsibility requires us to rethink assessment and teaching. Rather than focusing exclusively on detecting AI use, there is growing value in designing learning that makes judgement, reflection and process more visible.
These approaches (such as vivas, iterative drafting processes and collaborative problem-solving activities) are not new. In many ways, good AI-aware assessment looks remarkably like good assessment.
Our goal should not be to produce graduates who simply know how to operate AI tools. We have an opportunity to develop AI fluency: critical judgement, ethical reasoning, transparency and adaptability.
AI is neither inherently an enabler nor a barrier to inclusive education. It becomes one or the other through the choices we make as educators and institutions. The future question for higher education is not if AI belongs in higher education, it is whether universities can remain values-led and intentional in its use.