The Great Unlearning

Yalda Keshavarzi | Writer, Editor, Poet

Posted: Thu Nov 05 2026 | Last Updated: Thu Nov 05 2026
In an age of artificially abundant intelligence where a book can be summarised within seconds, our ability to read, write, or even think independently is under threat. Children’s author Katherine Rundell recently told the [*New Statesman*](https://www.facebook.com/NewStatesman/videos/almost-all-of-my-books-were-stolen-to-teach-ai-how-to-create-its-voicebut-what-k/1046105645096060/) that almost all her books had been stolen to teach AI how to create its voice. Describing her experience in her own words: “I find, at a purely personal level, deeply greasy and unpleasant, the sycophantic vile little voice of AI.” Firstly, it has to be said that AI could never articulate itself with quite so much fervor and I would urge everyone to listen to her interview. Secondly, there’s a grotesque irony in an industry extracting the intellectual labour of writers to build machines that encourage us to bypass the work of reading and thinking for ourselves. The AI overlords promise to [democratise knowledge](https://www.forbes.com/councils/forbestechcouncil/2024/06/25/ai-and-the-democratization-of-knowledge/), making information more accessible than ever. But what happens when having an answer becomes a substitute for understanding it? What does this mean for our capacity to question the people and institutions that hold power? ## **Too many shortcuts** Researching, wrestling with contradictory ideas, struggling to articulate an argument, and repeatedly revising it is how we learn to think. AI provides a shortcut through much of that process, producing in two blinks of an eye what might’ve take days (or months) of intellectual effort. Somewhere between that excruciating first draft and the euphoric final full stop, we develop patience and confidence in our judgement. That mental exhaustion is proven not to be in vain. In 2025, [researchers at MIT](https://www-prod.media.mit.edu/publications/your-brain-on-chatgpt/) studied the effects of using ChatGPT for essay writing, comparing participants who either wrote independently, used search engines, or relied on AI. Those using the addictive chatbot showed the weakest brain connectivity during the task and even struggled to recall passages from their own essays. Perhaps even more concerningly, when some participants who had used the large language model (LLM) were subsequently asked to write without it, their brain connectivity remained lower than those who had worked independently throughout. While the study was relatively small, its findings raise important questions around our retention of independent thought when we repeatedly outsource the process of thinking. There’s a habit to reduce illiteracy to the inability to read words on a page. But there’s a deeper kind of illiteracy at risk, such as being able to consume information without the skills to interrogate it or simply being able to spot when someone's talking nonsense. As we become accustomed to instant answers, I fear for our patience for murky questions and even murkier paths. ## **Who gets to think?** The consequences of this shift are unlikely to be shared equally. In a [*Guardian* essay](https://www.theguardian.com/books/ng-interactive/2026/aug/08/i-hate-what-ai-is-doing-to-the-minds-and-happiness-of-the-young-katherine-rundell-on-the-view-from-the-classroom), Rundell warned that AI could deepen educational inequality. As tech companies enthusiastically promote AI tutors as the future of accessible education, she raises the possibility of wealthy schools protecting traditional, human-led learning while poorer institutions face pressure to adopt cheaper automated alternatives. A future where the wealthy are taught to think, while the rest of us are taught to perfect a prompt is not one I'm particularly keen on. Of course, economic inequality already determines who has access to good schools, libraries, private tutors, and the time to study. Money doesn’t buy independent thought, but it can buy the space to develop it. It looks like AI will introduce another divide: between those encouraged to develop intellectual independence and those dependent on corporate technologies to do their thinking for them. Perhaps more concerning is handing over the development of young people’s education to some of the world's most powerful corporations whose primary concern is hardly the cultivation of enquiring minds. As much as they’d like us to forget, LLMs aren’t neutral. They’re commercial products shaped by corporate interests and decisions about which information to prioritise. Rather than democratising knowledge, AI concentrates greater power over [how knowledge is produced and interpreted](https://www.businessinsider.com/ai-is-handing-control-of-knowledge-to-big-tech-professor-2025-10). That power goes beyond what AI spits out to what it decides we should swallow. Across the US, schools have been removing thousands of books from their libraries. Unsurprisingly, the Trump era has targeted works dealing with race, sexuality, and political identity. Within this somewhat dystopian reality, according to [PEN America’s latest report](https://pen.org/report/book-bans-spotlight-2026/?stream=top), some school districts are now turning to generative AI to help determine which books should be removed. [Public records](https://apnews.com/article/book-bans-pen-america-stephen-king-6139f813e15c2b62ae4d27b06121ea41) revealed chatbots recommending the removal of books purely because other school districts had already banned them. They might have us think book bans and AI-generated answers are separate issues, but both point to greater controls over the ideas we’re exposed to. ## **AI resistance** Our ability to challenge injustice depends on more than having access to information. Whether we’re campaigning against new data centres, challenging corporate complicity in genocides, or questioning why our weekly food shop costs the same as a small holiday, understanding complex systems and questioning whose interests are served are essential skills. Ones that are developed through reading, discussions, disagreements, and the beautiful process of changing our minds. When we hand these processes over to AI we surrender the intellectual independence needed to challenge power. AI companies have built enormously profitable technologies on the accumulated work of [generations of artists, writers, and researchers](https://www.theguardian.com/commentisfree/2026/aug/05/anthropic-ai-destroying-books). As more of the world's wealth is concentrated in the hands of a select few, especially tech billionaires (and Elon Musk’s [undecided trillionaire status](https://www.forbes.com/sites/zacharyfolk/2026/10/13/elon-musk-no-longer-a-trillionaire-after-spacex-stock-slides-tuesday/)), we need to fight the concentration of power with intellectual wealth. --- ## [**Join Our March**](https://pulltheplug.uk/forms/sign-up-to-march-against-the-machines)