In June, academic staff from across the UK attended a BSA Sport Study Group workshop to explore how generative AI is transforming the teaching of the sociology of sport. The session addressed a fundamental question: if AI can instantly generate essays, conduct research, and provide feedback, what is the purpose of our assessments?

The informal online seminar, organised by Dr Jon Dart (Leeds Beckett University), Dr Jim Lusted (The Open University), and Dr Mark Doidge (Loughborough University), sought to move beyond the traditional assessment approaches that were focused on recall to explore how academic staff and students use Gen AI in their teaching, learning and assessments.

Innovative Teaching Approaches
Dr Michael Hobson (St. Mary’s University, London) began by sharing his experiences of using Gen AI. Among other things, Michael explained how he works with students to develop prompts that provide meaningful feedback on their work. He noted that while AI can help with reading academic articles, such as simplifying texts, translating, and extracting key points – it cannot replace the fundamental task and skill of (critical) reading.

A key concern raised was the potential inequality between students using premium (i.e., paid for)  AI tools versus free versions, alongside the risk that students might lose essential soft/durable skills when AI handles much of the analytical work.

Dr Conor Heffernan (Ulster University) presented his innovative role-play activity where students develop interview questions for AI personas (for example, governing body officials, recently retired players or referees). Students then analysed AI’s often superficial responses against published research, with the activity encouraging reflectivity and critical thinking while highlighting AI’s tendency toward ‘hallucinations’ requiring fact-checking.

Critical Reflections
Participants raised serious concerns about generative AI’s broader impacts and ethical issues. Beyond built-in biases affecting women, girls, and other minoritised groups, participants highlighted the often-overlooked issue of data ‘appropriation’, where Gen AI tools ‘steal’ information without consent. The significant environmental costs of using large language models (LLMs), which consume enormous amounts of energy and water, were also highlighted. It was noted that this critical understanding of AI should be an important component of future teaching to contextualise its use.

Participants also expressed their concern that, while AI tools might have the potential to reduce staff workloads, its use can also lead to significant increases in workload. In addition, some raised fears that AI adoption in HE may lead to further staff cuts and redundancies. It was also noted that large language models often perpetuate existing knowledge hierarchies that marginalise already disadvantaged communities.

Although participants had more questions than they started the session with, AI presents both opportunities and challenges that require thoughtful, critical engagement rather than wholesale adoption. This workshop highlighted the importance of ongoing collective discussions and sharing of practice around AI, which was particularly beneficial among colleagues who regularly teach the same discipline.