What are the consequences of letting AI read for us?
Naomi Baron, professor emerita of linguistics at American University, explores this issue.
Faculty Bio:
Naomi S. Baron is Professor Emerita of Linguistics at American University in Washington, DC. She is a Stanford PhD and has been a Guggenheim Fellow, Fulbright Fellow, and Visiting Scholar at the Stanford Center for Advanced Study in the Behavioral Sciences.
Author of eleven books, she has done research in a wide swath of areas, including child language acquisition, the history of English, writing as a form of linguistic representation, reading, and artificial intelligence. The emergence of email, mobile phones, and social media led her to explore how online and mobile communication reshape our linguistic interactions (Alphabet to Email, 2000, and Always On, 2008). Two subsequent books (Words Onscreen, 2015, and How We Read Now, 2021) drew on her empirical studies of how users view print versus digital screens as reading media.
More recently, Professor Baron has focused on the ways in which artificial intelligence is reshaping what it means to write and read. Who Wrote This? (2023) broached the challenges we face when AI becomes our writing or editing surrogate. Her new book, Reader Bot (2026), takes on the complementary issue of outsourcing human reading to a large language model.
Transcripts:
Since writing first emerged roughly 5,000 years ago, motivations for reading have proliferated. From legal declarations to religious creeds, epic poems to novels, research articles to emails, the texts we read invite us to connect with the words and thoughts of others.
Today, humans aren’t the only ones reading. As large language models (LLMs) permeate our lives, we increasingly lean on them to read on our behalf.
AI-as-reader can produce an impressive comparison of Mark Twain’s The Adventures of Huckleberry Finn and Percival Everett’s James. It can locate, summarize, and analyze scores of research papers. Human labor is reduced, but with what consequences? The question is increasingly pressing, given worrisome declines in how much both children and adults are reading of their own volition. Documenting a related trend, my research with Anne Mangen found that 49% of the US university faculty we surveyed were assigning less reading than in the past, with 34% indicating this was because students were ignoring the assignments.
But the value of reading far supersedes completing academic or professional tasks.
My research on AI and reading probes mental consequences of outsourcing reading, along with benefits from reading ourselves. Both experiments and self-reports indicate diminished cognitive effort and potential deskilling if AI reads for us. When doing our own reading, we build human connections, perhaps leading to empathy. We relax, escape, or peer into the minds of others. Reading encourages us to contemplate whom we want to emulate or what values we might aspire to. Coming-of-age novels (including Huckleberry Finn and James) enable us to vicariously share in – and learn from – the protagonist’s journey.
Turning to AI for some reading tasks, some of the time, is a principled choice. However, in our AI-saturated world, it behooves us not to ignore the personal rewards human reading can bring.
Read More:
[Stanford University Press] - Reader Bot - What Happens When AI Reads and Why It Matters
[Stanford University Press] - Who Wrote This? - How AI and the Lure of Efficiency Threaten Human Writing
[Taylor & Francis Group] - Student Perceptions and Practices When Reading in Print and Digitally
[Duke University Press] - Doing the Reading: The Decline of Long Long-Form Reading in Higher Education
