Starting, scrapping, and rebuilding AI literacy at Sussex #LILAC26

This is my first blog post for the 2nd day of the LILAC information literacy conference is From workshop to workshop: Starting, scrapping, and rebuilding AI literacy at Sussex presented by Nicholas Heavey (University of Sussex, UK). The abstract is here
Heavey said that initial initiatives felt a bit piecemeal, so they first identified a framework to help map skills across the curriculum. They decided to use UNESCO's framework for AI competency. This has 3 progression levels and 4 competency aspects. This helped them to see where gaps were, what they needed to cover and what they didn't. As this framework was designed for schools, they also mapped it to the UK's QAA HE framework, which gave a roadmap and helped to break this down for session aims and learning outcomes. Also it gave legitimacy when they went to talk to academics.
Heavey said that the first workshops didn't go quite as planned but gave useful insights into how students were actually using AI & what they needed. The library team went on to a dialogic and problem based approach, e.g. asking learners iteratively using prompts to see what happens and develop effective prompting techniques.
Looking at what worked at what didn't - early workshops were too technical, overloading learners with detail they didn't need. Therefore they turned to a more experiential learning approach, with more time to reflect. Secondly, they noticed that learners were anxious about AI e.g. not wanting to use it in a way that would make them discredited. Therefore it important to have open conversations, not being the assessment police and also not overstating the case against AI. Thirdly they have introduced more playful elements, e.g. using lower stakes activites/ tools to encourage experimentation. Two of their main sessions are: Questioning & prompting and Chatting and searching.
Heavey identified that sessions are structured around the brain vs AI (so thinking about what just uses the brain, what uses LLMs, how can you effectively use LLMs to assist) "LLM and research tasks emphasise the value and importance of researching and writing". The focus is on learning and understanding and they ask the learners to think of themselves as researchers. They explain key concepts through analogies and metaphors, which can be used to stimulate discussion. They also remind learners that they are responsible for any AI generated material. They ask people to reflect on whether the response from AI is changing the question they are asking - whether it actually answers the question they are posing.
The team use the CLEAR framework for AI prompts (example explanation here). Heavey presented a diagram to do with tool choice with more creative at one end and less creative at the other.
They use a three level model - introduction; Critical AI skills; and (not developed yet!) Advanced AI applications. They have also collaborated with teaching faculty (which sounded like a good collaboration with co-design and co-teaching), and gave an example of collaboration with Law where students took on the role of trainee solicitors (they undertook an authentic task, evaluated LLM outputs and reflected on professional ansd ethical implications). This was embedding traditional and AI research skills together.
In conclusion, Heavey saw this as an opportunity, as academics and learners are looking for support, and librarians can show how they can help.
Photo by Sheila Webber: not AI generated - magnolia tree in Sheffield Botanic Gardens, March 2026.