"Normal Service Won't be Resumed" - A Reflective Case Study of AI Literacy, Assessment and Academic Integrity

<p xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" class="first" dir="auto" id="d550834e85">This reflective case study reports on a three-year programme of AI literacy led by Learning Design and Curriculum Development Advisers at Newcastle. Initially conceived as a cautious technology-adoption webinar, the work was often redirected towards conversations about risk and ethics at both societal and sector levels. Participants routinely raised concerns about exploitation in training data, environmental damage, copyright, equity, cognitive decline and algorithmic bias. Practically, we were required to foreground academic integrity, with assessment positioned as the primary safeguard. This placed academic judgement, assessment design, and policy and practice under particular scrutiny in the context of rapidly evolving AI tools.Across the period there were more than 1,000 workshop attendances, including around 15 within discipline or school-specific workshops and university-wide open sessions attended by some 200 colleagues. Workshops were iteratively co-developed with participants, guided by a shared goal of surfacing and circulating emerging practice. It quickly became clear that no single department or unit could offer definitive answers on AI use, risk or ethics, and that questions about policy, data use and the vulnerability of disciplinary practices to automation routinely exceeded what could be resolved in a single session.The programme also revealed that AI did not feel like “just another” episode of digital adoption. Excitement and curiosity about automation co-existed with anxiety, anger and perceived erosion of professional identity, alongside hopes for inclusion, access and workload relief. These divergent emotional responses shaped how participants interpreted evidence about AI risks and benefits and their willingness to experiment.This paperargues that AI literacy should be treated as a substantive educational goal in its own right and as a necessarily dialogic, ongoing practice. It concludes by outlining principles for AI literacy interventions that make value conflicts discussable, recognise the emotional dimensions of AI adoption and support informed, context-sensitive institutional decision-making.