Artificial intelligence and educational techno-panics: a lesson from history

The rapid development of artificial intelligence (AI), and well-known tools such as ChatGPT in particular, have led to a moral panic about its deleterious effects on academic standards in higher education. This conceptual paper argues that AI in education needs to be understood from a historical and sociological perspective as a socially constructed ‘moral panic’ (Cohen, 1972) about the role of technology in society (Nye, 1996) and specifically in respect to authorship. Widespread concerns have been expressed that AI will increase student cheating, spread misinformation, and potentially undermine the trust and reliability of academic research as well. It also feeds the trope that students cannot be trusted to learn and are more likely to cheat than academic faculty (Macfarlane, 2017) illustrating Winner’s (1980) argument that technology reinforces authority and power relations. There is a long history of moral panics connected with technology and academic standards which in the 1990s and 2000s were centred on the spread of the internet and associated worries that it would facilitate student cheating. AI may represent a further step change in technology but historically it represents an evolution as a tool for authorship building on existing resources or ‘cheat functions’ in relation, for example, to spelling and grammar, online thesauruses, reference management software, web-based citation generators and the resources of the internet more generally. Even paper-based tools such as encyclopaedias and dictionaries have long been used by authors without a moral panic that authors may be cheating. While predictive text is AI generated its common availability as a default option on contemporary phones and laptops means that authors who associate plagiarism with a ‘cut and paste’ culture may not understand that they have even used AI. This may be conceived more as a continuum from human to machine authorship, akin to Walker’s (1998) plagiarism scale. The current moral panic centring on machines as co-authors also distracts attention from a perhaps more significant scandal concerning human co-authorship fraud. Here evidence suggests that almost a half of authors listed in co-authored papers fail international guidelines with respect to human co-authorship (Sauermann & Haeussler, 2017). From a sociological and historical perspective, AI needs to be understood as the latest ‘folk devil’ to emerge onto the educational landscape and not the only contemporary threat to authorship integrity.