If you were to create a tool intended to support authoritarian control, it would bear a striking resemblance to artificial intelligence. Its capacity to distort our shared understanding and generate disinformation is unmatched in human history. With AI, there's no need for legions of writers and artists to disseminate propaganda. It facilitates extensive surveillance, placing constraints on those who oppose you. Its lethal autonomous weaponry can streamline and normalize remote killings. Moreover, it exacerbates inequality, already driving an alarming concentration of power and wealth into the hands of a few individuals whose views on the intrinsic value of human life are deeply troubling.
Artificial intelligence holds the potential for incredible advancements. It could transform healthcare, accelerate the search for cancer cures, support individuals with disabilities, change our comprehension of data, increase accessibility to legal resources, enhance early warning systems for disasters, and reveal previously hidden truths about our planet's ecosystems. However, if harnessed by nefarious actors, such benefits will become irrelevant.
For aspiring despots, large language models (LLMs) like ChatGPT, Claude, and Google’s Gemini have disrupted centuries of educational norms, undermining our ability to engage in critical thought essential for pursuing justice. As a children's author and a fellow at Oxford University, I frequently encounter the influence of generative AI on the youth. I've observed its intrusion into their lives and the urgent need to emphasize how vital reading and literature are to our identity and democratic society.
There are numerous aspects of AI that I fundamentally despise. The technology's capacity to create counterfeit individuals—artificially generated personas—threatens the very fabric of civilization. Many LLMs have been developed through unethical means, capitalizing on human labor without proper compensation. The environmental toll of AI is alarming; by the end of this decade, the energy demands of AI-optimized data centers are predicted to surpass those of all of Japan. Furthermore, AI could potentially push Earth’s ecosystems into irreversible disarray.
Of greatest concern is AI’s impact on the mental well-being of young people. After engaging with thousands of children in schools, I am increasingly convinced that if we truly wish for justice to guide our society, AI should not play a role in the lives of children and adolescents. It threatens to erect barriers that inhibit their potential for critical thinking, experimentation, resilience, and creativity—ultimately obstructing their paths to freedom. It risks undermining genuine learning before it even begins.
Proponents of AI claim it will enhance educational efficiency. But is that really the aspiration we have for our children? To lead an "efficient" childhood? Research consistently shows that when aid from AI enters the educational sphere, students tend to give up on tasks earlier. One study suggested that even after just ten minutes of AI-assisted problem-solving, those who lost access to AI performed worse and gave up more frequently than those who never used it. An MIT study also indicated that students using ChatGPT exhibited lower brain activity and underperformed in linguistic and behavioral measures. This leads to cognitive offloading and, ultimately, cognitive surrender—making us forget what we once found possible.
Children have long exhibited a stronger connection to physical books than digital screens; a tactile experience is integral to understanding. Yet, we are now witnessing a surge in educational technology that pushes learning online. It’s common for my teenage goddaughter to inform me that nearly all her classmates rely on ChatGPT to complete their assignments. The polished work produced by AI often gets misperceived by students themselves as a sign of their own mastery. When the output no longer reflects a child’s comprehension, it exposes them to the allure of speed and fear of appearing inadequate, compromising their intellectual clarity.
Those in the edtech sphere advocate that AI will liberate educators and create individualized learning opportunities for disadvantaged students. While AI-driven education may appear cheaper than hiring teaching assistants, history shows that institutions tend to opt for the more affordable option—even if it's of significantly lower quality. In reality, elite schools may gravitate towards non-tech-based education models, while under-resourced institutions will likely depend on LLMs more heavily.
The idea that AI leads to efficient learning leads to unsettling questions: Is an efficient childhood truly desirable? Should we prioritize efficiency in storytelling or artistic expression? A competent teacher does more than just recount historical facts; they guide children through understanding human emotions, constructing arguments, engaging an audience, and instilling a love for learning. A machine can only teach them to mimic computer-like behavior.
Continuous engagement with technology characterized by pseudo-human interactions poses grave risks for young minds. My concerns deepen upon reflecting on tragedies like that of Sewell Setzer III, a 14-year-old who took his own life after being urged by a chatbot with an unsettlingly human-like voice.
A common counterargument to my apprehensions about AI in schools is the notion that students need to learn how to use these tools to stay relevant. However, today’s technology will evolve dramatically in the near future; teaching children how to wield current AI is akin to exchanging their intellectual independence for an obsolete tool. In a conversation with the president of a prominent tech company, he stated he would not teach his children to use AI because it would prepare them for nothing. Instead, he would focus on imparting knowledge related to the humanities—understanding who we are and what we’re capable of creating. An observation made by the Financial Times further highlights a trend where employers now value humanities graduates, as “AI-native” STEM alumni often enter the workforce with disappointingly superficial ideas.
In higher education, faculty members have already noted the detrimental effects of a decade dominated by smartphones on the intellectual development of students. Students frequently arrive having read significantly fewer books, finding it harder to analyze complex texts. With the emergence of LLMs, the situation has only intensified. It has always been possible to generate essays without fully engaging with the source material; however, now we confront the unprecedented reality of students fabricating essays from AI without reading any component of what this AI-generated text has produced.
Professor Paul Musgrave of Georgetown University emphasizes the decline in student engagement with reading and assignments, noting that the difference in classroom discussions stemming from disparate levels of reading completion is profound. Students’ reluctance to work is unmistakable, yet the dangers of AI temptations are ever-present, leading some to feel penalized for their honesty. I encourage students to view their education through the lens of an Olympic athlete. While taking shortcuts—like riding a bus—might seem appealing for a quicker outcome, the true value lies in the training and strength developed.
In the context of learning, a child’s essay on a literary work holds little intrinsic worth if it’s the result of AI, as the true essence lies in the knowledge and critical thinking cultivated during the process of writing. Collaborating with AI for educational tasks is akin to relying on transportation rather than pushing beyond one's limits.
We can’t be assured that the work students submit is genuinely their own. Colleagues at universities in the UK express confidence in distinguishing AI-generated submissions, but I believe it’s increasingly difficult to do so based solely on text. A friend teaching mathematics remarked that some essays still contain AI’s upbeat “Happy to help!” line. While identifying poorly used AI isn’t overly complex, a well-crafted, edited AI-generated essay could easily be mistaken for original work, especially amid the substantial pressures young people face today.
Academics, especially those in fields like law where misinterpretation has dire consequences, are rightly distressed by governmental inaction regarding the threats posed by AI to our capacity for independent thought—one of the cornerstones of democracy.
What solutions exist? I’m skeptical that the commonly proposed methods—like oral exams and frequent tests—can adequately replace the deep, transformative process that essay writing embodies. Writing is not simply about transcribing thoughts; it’s a method of refinement, morphing a rough idea into something coherent. As E.M. Forster said, “How do I know what I think until I see what I say?”
Although I don't have a definitive answer, I believe we must resist passivity. Young minds are extraordinary, and squandering their potential would betray the trust placed in the older generations. Integrating AI into education without acknowledging its detrimental impact on developing intellect is a mistake. As we witness AI-generated personas infiltrating our media, we cannot afford to equip young people with hollow knowledge.
Many young individuals express their dissatisfaction with the rise of artificial intelligence in their lives and seek ways to resist. I advise them to view knowledge and creativity as powerful forms of resistance. Engaging with books—and discussing their experiences within a community—can fuel their fight against technology-driven instability.
I frequently emphasize the importance of reading for enjoyment and exploring the inherent joy of literature, highlighting its essential role in education. However, it’s crucial not to confine our definition of pleasure. The learning journey can often be arduous; true breakthroughs are rare and require effort.
We must also encourage youth to read not only for enjoyment but in pursuit of justice and to strengthen their critical thinking and creative abilities. By immersing themselves in human literature, they affirm their humanity and cultivate the mental resilience necessary to combat technological encroachment. Allowing machines to dominate our thought processes jeopardizes the very essence of civilization.
Challenging days lie ahead. To confront those wielding immense financial resources, we’ll need equally substantial intellectual capital. We require individuals who are strong-minded, knowledgeable, and perceptive enough to recognize the realities of our situation in real time. While enjoyment in reading is undeniably important, we must also prioritize reading as a means of empowerment.
