Rethinking Generative AI in Human-Computer Interaction: Review of Affordances, Constraints, and Implications

Generative Artificial Intelligence (GAI) transforms our technological interactions, including new capabilities and concerns about biases and misuse. In the field of human-computer interaction (HCI), previous research has investigated generative AI in relation to human-centred AI, user trust, user experience, design work, co-creativity, and user personas. This study applies the theoretical lens of affordances and constraints to ask the question: Which affordances and constraints of generative AI can be identified in human-computer interaction research? The study employs a scoping literature review approach to collect data from the Web of Science Core Collection databases. The query string combined keywords, such as “generative”, “artificial intelligence”, and “human computer interaction”, with Boolean operators AND and OR. Inclusion and exclusion criteria were used in the screening of 156 identified articles, from which a total of 37 were selected for inclusion in the study. An initial categorization matrix, based on the theory of affordances, was used to conduct a deductive thematic analysis. The analysis followed the guidelines for thematic analysis suggested by Braun and Clarke. The investigation identified seven key themes, with included sub-themes, illustrating the varied applications and potential effects of generative AI. The seven key themes are: 1) improving algorithms, 2) collaborative work, 3) education support, 4) truth issues, 5) biases, 6) ethical considerations, and 7) consequences for job market. The study further highlights the importance of considering contextual differences and short-term and long-term consequences when applying GAI technologies, as well as ethical considerations, such as ethical and legal accountability. The paper concludes with a novel conceptual model for affordances and constraints of generative AI, informing future research, guiding stakeholders’ use and implementation, and providing design recommendations for generative AI systems across various sectors.