Artificial instrumentalism? A critique of AI acceptance and AI anxiety research

Abstract The ongoing proliferation of AI technologies has created a need to understand and model how people relate to, accept, and feel about AI. Among the research traditions that attempt to capture how individuals navigate these changes are those centred on the concepts of “AI acceptance” and “AI anxiety”. This perspective paper suggests that the typical use of these concepts often rests on instrumentalist perspectives of technology that are especially problematic in the context of complex sociotechnical transformations and ethical AI adoption. Firstly, AI acceptance, when based on the Technology Acceptance Model (TAM), presupposes technology acceptance as inherently rational and positive and focuses disproportionately on overcoming barriers to use, which limits its capacity to capture how people relate to AI. Secondly, the concept of AI anxiety has been used to frame critical and hesitant perspectives of AI as nonrational. By comparing these concepts and analysing their historically instrumentalist underpinnings, it is argued that both research currents require further theoretical examination and critical engagement. As contextual examples, I examine AI anxiety and acceptance as guiding concepts in select contexts, mainly education. I argue that research should engage with AI adoption as sociotechnical change where ethics, uncertainty, rationality and nonrationality interact, impacting use, non-use and sociotechnical reconfiguration.