The Greasy Little Voice: Why Katherine Rundell’s AI Warning Matters for Our Future
In a searing and beautifully argued essay, author and Oxford fellow Katherine Rundell puts her finger on a crisis unfolding across classrooms, lecture halls, and quiet corners of daily life. The threat from generative artificial intelligence is not merely that it spits out uninspired "slop" or automates routine tasks. It is that AI, delivered with what Rundell strikingly calls its "greasy little fluent voice," risks quietly gutting our capacity for critical thought, stamina, and intellectual independence.
Rundell speaks from the front lines. Working continuously with young people, she warns that large language models—built on what she rightly describes as an "illegal pillage of human labour"—are lowering an iron curtain between developing minds and their true potential.
Her critique raises urgent concerns that hit at the very heart of everything we stand for at Tax the AI:
Cognitive Surrender: When we outsource thinking to software, we do not gain efficiency—we lose neural capability and moral stamina.
The Ed-Tech Class Divide: Wealthier families and private schools will pay for human teachers and physical books, while state-funded schools are left with cheap algorithmic substitutes.
The Shift of Power and Wealth: Unchecked tech expansion concentrates immense power in a tiny handful of hands while imposing huge environmental and social costs on the rest of us.
1. Cognitive Surrender and the Illusion of Mastery
Rundell points to compelling research detailing the physical and mental cost of relying on LLMs. Studies reveal that after just ten minutes of AI-assisted problem-solving, individuals perform worse and give up far more quickly once the software is removed. An MIT study showed that test groups using ChatGPT exhibited the lowest overall brain activity, underperforming across neural, linguistic, and behavioural metrics.
This process is not a harmless shortcut; it is genuine cognitive surrender.
When a student relies on AI to write an essay, they are not collaborating with a helpful partner. As Rundell writes: "you can't collaborate with AI in the same way you can't collaborate with a bus." The essay was never meant to be a commercial end product; its value lay entirely in the mental resistance built while wrestling with ideas to create it. When young people mistake an algorithm’s smooth prose for their own mastery, we strip them of the very scaffolding needed to build independent minds.
2. A Two-Tier Educational Divide
One of Rundell’s most acute warnings focuses on the false promises of educational technology. Proponents claim AI will democratise learning and free up educators. In reality, economic pressure will force a far darker outcome:
Privileged institutions will follow the example of Sweden and Norway, swerving back towards paper books, handwriting, and human-led instruction.
Underfunded state schools will face immense pressure to replace teaching assistants and specialist staff with low-cost software subscriptions.
This creates a deeply unequal landscape. The wealthy will purchase genuine human mentorship, while working-class children are handed over to automated chatbots—widening societal divides under the banner of modern "efficiency."
3. The Shift of Power and Unpriced Externalities
Rundell does not shy away from the macro-political stakes. If you set out to design a tool to facilitate authoritarian rule, she notes, it would look remarkably like generative AI: an engine for mass disinformation, surveillance, and remote control.
At the same time, this technology drives an astonishing transfer of wealth from ordinary working people to a handful of corporate executives. It demands staggering resources—with global data center electricity consumption rising sharply—while privatising the profits and passing the environmental damage onto local communities.
Rebalancing the System: Connecting Rundell's Warning to Tax the AI
Katherine Rundell’s plea reinforces the core mission behind our work on this platform:
Taxing Capital Windfalls over Human Labour: Generative AI relies on the uncompensated work of millions of writers, artists, and workers. When corporations use that stolen labour to replace human jobs, an explicit AI and Robot Tax is essential to capture that windfall and level the economic playing field.
Funding Human Education: Revenues raised through automation levies must directly fund human teachers, smaller class sizes, and classroom support—ensuring every child receives human mentorship rather than an algorithmic script.
Pricing Environmental Costs: The massive water and energy footprint of training data centres must be properly taxed to curb wasteful technological expansion and protect our shared environment.
Taxing AI is not about halting useful technological breakthroughs in medicine or environmental science. It is about claiming a necessary social dividend from an extractive business model. If tech firms are permitted to displace human effort without contributing fairly to the society that made their software possible, democracy itself suffers.
Reclaiming Knowledge as Power
Rundell leaves us with a clear call to action. Resistance begins with protecting our intellectual independence. We must encourage young people to read—not simply for passive entertainment, but for strength, stamina, and freedom.
To build a fair economy and protect democratic choices, we need tough-minded citizens who can analyse, question, and see through smooth illusions. We cannot stand by while a generation is sent out into a challenging world equipped only with counterfeit knowledge.
It is time to hold the architects of unchecked automation accountable—economically, socially, and politically.
