Learning designers and AI: Navigating values, ethics, and future capabilities

Kay Harrison, Sangeetha Kirsnan, Simone Poulsen, Tina Barclay and Monica Vahl examine AI adoption among learning designers and other workers supporting teaching and research. Their ASCILITE 2025 paper studies professional judgement and institutional support, rather than training an AI model to adopt respondents’ values.^[1]

What the survey measures

An anonymous 5–10-minute online survey circulated through email and professional communities. Respondents self-selected; the authors report 49 completions from 81 alongside a 74.24% response rate. Those figures do not agree arithmetically, so the reported percentage is not a reliable completion rate. The analysis combines descriptive statistics with exploratory thematic interpretation, principally organised using existing frameworks.^[1]

Respondents selected multiple values rather than ranking a single collective objective. Curiosity appeared in 84% of personal-value responses and 43% of perceived institutional-value responses; ethical responsibility appeared in 71% and 53%, respectively. These compare individuals’ reports about themselves with their impressions of institutions, rather than an independent assessment of institutional priorities.^[1]

Support and ethical judgement

Table 2 makes support more concrete. Of 49 respondents, 29 reported some professional-development opportunities and nine reported a lot; technical support was less available, with 11 reporting very little and nine none. These categories describe perceived provision, not the effectiveness of training or tools. The authors also report a split between confidence in applying ethical considerations (53%) and lower confidence (47%).^[1]

The discussion distinguishes learning a tool from exercising judgement about its outputs and educational purposes. Proposed responses include checking outputs and bias at design checkpoints, preserving collaboration with colleagues, and connecting existing assessment and facilitation skills to AI work. These are recommendations arising from the survey and literature, not interventions tested by this study.^[1]

Scope and alignment relevance

The two-week collection period, small self-selected sample and absence of respondent location data limit generalisation. The authors acknowledge broad survey questions and interpretive bias. The study does not measure causal effects of AI adoption, student outcomes, or whether the proposed professional-development changes work.^[1]

For Social Alignment, its contribution is to show why workplace adoption cannot be reduced to individual enthusiasm or institutional instruction. People who help implement AI also need resources and a voice in deciding how ethical expectations apply to their work. Choosing whose judgement counts remains an institutional question alongside technical capability.