Gabriella Cagliesi is a Professor in the Department of Economics and Teaching and Learning Lead at the University of Sussex Business School. Since joining Sussex in 2019, she has championed innovative and inclusive teaching, earning recognition for initiatives that close attainment gaps and enhance student experience. 

With a PhD in Economics from the University of Pennsylvania and over thirty years of teaching experience, Gabriella is research-active in applied international macro-finance, applied behavioural economics, and empirical studies on labour markets and educational choices and policies. She also collaborates on projects that support widening participation and enhance student outcomes.  

This case study illustrates how Gabriella integrates generative AI into her teaching, encouraging students to use AI as a study tool and engage with it critically and creatively.  

1. How do you bring generative AI into your teaching? 

I integrate generative AI through a customised Chat GPT environment, rather than students using open platforms. This involves creating a closed system where I upload teaching materials, define the AI’s role, and set clear boundaries on what it can and cannot do. In seminars, students work in groups using this custom AI tool, and I emphasize ethical use and the risk of hallucinations of even a closed and bespoke AI platform. AI plays different roles across the teaching sessions, such as a teammate (acting as a devil’s advocate), tutor, Socratic teacher, simulator, or podcasting assistant. After each activity, students reflect on AI’s role and submit their interaction logs, which I review and provide feedback on. 

2. What approaches do you use to integrate AI into your teaching, and why have you chosen them? 

I use three main approaches: 

• Interactive learning design: AI is embedded in classroom activities to encourage experimentation, scenario testing, and critical thinking. 

• Teaching material development: AI helps create study guides and summaries of teaching, accelerating routine tasks while maintaining transparency with students. 

• Assessment integration: AI is incorporated into assessments through synthetic datasets and reflective tasks, requiring students to critique prompts and evaluate AI’s limitations. 

These approaches were chosen because they align with my discipline (economics), promote higher-order thinking skills, and prepare students for real-world applications of AI. 

3. What impact has AI had on student learning, curriculum design, or academic practice? 

AI has significantly enhanced student engagement and understanding. Students report that simulations and visualizations help them grasp complex concepts better than formulas alone. It has improved critical thinking, as students learn to question assumptions and evaluate AI outputs. Curriculum-wise, I redesigned assessments to include AI-enabled tasks and reflection components, shifting focus toward interpretation, reasoning, and AI literacy. For me professionally, this work has led to invitations to present at conferences and collaborate across departments, fostering broader pedagogical discussions. 

4. Looking back, what would you do differently? 

Initially, I introduced AI only during seminars, but I now realise the value of pre-class integration. Allowing students to explore AI before sessions would have deepened engagement. I would also have focused earlier on student learning through AI, rather than policing its use. Designing prompts that encourage reflection, and reasoning has proven more effective than simply controlling access. 

5. Three practical tips for fellow academics: 

1. Build your own AI literacy and confidence: Experiment with tools to understand their capabilities and limitations. Confidence in using AI translates into better student experiences. 

2. Shift from content delivery to challenge design: Create tasks where AI supports but does not replace human judgment. Clearly define acceptable uses and disclosure requirements. 

3. Use AI to deepen reflection, not replace it: Incorporate reflective activities where students critique their own reasoning and AI’s output. This fosters metacognition and critical engagement.