Concerns are mounting that generative artificial intelligence is contributing to a weakening of reading skills and an impoverishment of close attention, even as technology firms scan and digitise millions of printed books to train their models.
AI, reading and cultural loss
The critique is twofold. First, widespread use of AI to summarise books, articles or long email threads is said to be encouraging “intellectual sloth” — a diminished inclination to read slowly, parse language and savour nuance. Second, companies building AI assistants are reportedly acquiring vast quantities of printed material, unbinding volumes and converting pages into digital files, an approach described by critics as a violent assimilation of literary culture.
Those raising alarm point to the risk that the routinised delegation of reading and synthesis tasks to algorithms could corrode abilities that underpin empathy, detailed comprehension and imaginative immersion. The word “artificial”, they argue, captures an element of contrivance and commercial calculation — a technology-driven shortcut that substitutes immediacy for effort.
“For it would only be a capitalist like Sam Altman who would say ... that parents could ask ChatGPT to create a podcast they could listen to every morning centred around their children’s activities.”
The quote, cited in the original commentary, was deployed to illustrate objections to an expanding role for AI in intimate, everyday relationships and in functions traditionally performed through direct human interaction.
Anthropic and the scanning of books
The concerns extend to specific industry practices. The parent company of the AI assistant Claude, Anthropic, has been described as acquiring large collections of books — both old and new — which are then unbound and scanned in order to convert their pages into machine-readable text for training purposes. Critics say this process treats books as raw input to be consumed into a model rather than as objects of cultural value or as material that merits careful stewardship.
That combination — a cultural shift away from prolonged reading and an industrial-scale ingestion of print material — sits at the heart of the current debate. Advocates of deep reading warn that summarisation technologies can shortcut the cognitive work of interpretation and context-building that reading long-form material fosters.
- Reduced reading practice: Reliance on AI summaries may lessen people’s willingness to read full texts.
- Cultural processing: The unbinding and scanning of books for AI training raises questions about preservation and respect for literary works.
- Interpersonal effects: Critics argue that substituting algorithmic outputs for human conversation could weaken authentic relationships.
Those concerns are not presented here as definitive findings; they reflect a critical perspective circulating among commentators who value books and close reading. The discussion touches on technology, education, intellectual life and the ethics of data acquisition.
| Issue | Concern |
|---|---|
| Summaries replacing reading | Loss of nuance, context and empathetic engagement |
| Mass digitisation of books | Treatment of books as data rather than cultural artifacts |
| AI in family life | Risk of reducing direct human interaction |
For readers and households this debate has practical implications. It raises questions about how and when to use AI tools to summarise or interpret content, and whether such tools should supplement rather than replace direct engagement with texts. It also prompts consideration of the policies and transparency around how companies acquire and process copyrighted or out-of-print material for model training.
As the use of generative AI becomes more embedded in everyday routines, the tension between convenience and cultural practice is likely to intensify. At stake is not only the future shape of reading habits but also how society chooses to treat the literary record while building the next generation of technologies.
