Beyond Chatbots: Artificial Intelligence, Digital Reading Platforms, and Automated Pedagogy

• In Event: Artificial Intelligence (AI) and Literacy Education

Abstract

Current debates over whether and how literacy education should respond to innovations in artificial intelligence often position it as a novel phenomenon, or if not new, one primarily related to the disruptive effects of platforms like ChatGPT. But is the influence of AI on literacy education really all that novel, and might it manifest in ways other than chatbots? The answer to such questions may well depend on how one conceptualizes artificial intelligence, a term deployed across contexts with differing meanings and purposes (Kaspersen & Leins, 2022). For some, the term evokes AI's depictions in science fiction, where more or less computationally conscious agents pose solutions, problems, and often peril for humanity. For others, it denotes an array of computational techniques (e.g., machine learning), a subset of which are responsible for the large language models animating so much of the contemporary enthusiasm and anxiety about AI, as well as a host of more familiar technical processes (e.g., recommendation algorithms).

The argument developed in this paper is that when AI is understood as a set of computational practices, not as a nebulously autonomous mind, it becomes easier to appreciate how it shaped literacy education long before OpenAI deployed the ChatGPT platform, and continues to do so in ways that may be obscured by the current hype cycle. To develop this argument and explore its implications, I investigated the role AI plays across digital reading platforms (DRPs) (e.g., Accelerated Reader) by examining the kinds of promises they make to students, parents, and educators through the technical and discursive power of AI. The project builds on scholarship investigating the imbrication of digital platforms and literacy (e.g., Nichols & LeBlanc, 2021; Apps et al., 2023) in an effort to understand how AI fits within a broader regime of data practices reconfiguring how literacy is lived and learned.

Methodologically, the study has been guided by Light et al.’s (2018) walkthrough method for investigating apps, which I have adapted, following Apps et al. (2023), to the study of DRPs. I began by curating a set of keywords from scholarship on AI in education (e.g., Cardona et al., 2023; Williamson et al., 2023). I then used the keywords to conduct a systematic site search and walkthrough of a selection of widely used DRPs in the U.S., including Accelerated Reader, Lalilo, Epic!, I-Ready, Prodigy Reading, and I-Station. While searching, I tracked each term’s occurrences and examined their discursive effects, which allowed me to develop an emerging sense of how and to what ends DRPs deploy AI in literacy education.

Broadly, the study demonstrates how the growing influence of AI technologies on literacy education is indebted to the material infrastructures afforded by DRPs. It also highlights how DRPs use AI to facilitate the expansion of specific theoretical commitments (e.g., science of reading), thereby automating pedagogy in ways that may narrow the purposes of reading away from situated, communal forms of reading and towards a hyper-personalized, datafied literacy agenda.

Author

• Bradley Robinson, Texas State University