Pam McKinney here, continuing the live-blogging from the ECIL conference. This presentation was from Monika Krakowska, Magdalena Zych from Jagiellonian University, Krakow, Poland, the lovely location for the ECIL conference in 2023. Monika and Magdalena spoke about their research to understand ethical IL literacy in Higher Education. They wanted to diagnose LIS students' practices in working with Gen AI, and the gaps in students' real-world practices and the ACRL framework for IL. The research questions focused on how students understand authorship and disclosers of AI text production, how they assess the credibility of AI-generated responses and how criticality manifests in AI use. They used selected ACRL frames: information has value, authority is constructed and contextual; research as inquiry and searching as strategic exploration, and tried to understand how they could be reinterpreted in the context of GenAI.
84 LIS students at various levels of education took part in the research, they had to choose one of three tasks with ethical AI use dilemmas, modelling three roles: science fiction writer, information broker and podcast creator (e.g. analyse the consequences of using AI in video games). The team analysed the chat log from students' interactions with ChatGPT, and analysed this thematically, using the ACRL frames to support coding. Students mostly did not acknowledge the authorship of ChatGPT, did not question the credibility of the information provided by the AI, and did not identify bias. There was little iterative, purposeful prompting and searching, the prompt engineering was not very successful. Simple single-step requests/prompts dominated the chat logs, and usually focused on the format of the responses rather than the content.
In terms of the ACRL framework, they found that there was a need to examine AI use and how the ACRL framework could be updated and developed. Coming from the research are three central points: the need to recognise non-human contribution, bias and hallucination literacy is needed and prompt engineering competency is needed. The research took place in one university and there is an appetite for extending this to other universities. Future-ready IL must be AI-aware, AI-critical and AI-responsible.
Photo: rainy Monday morning in Bamberg (Pam McKinney)
No comments:
Post a Comment