Imagine you are a student in a U.S. public school classroom in the year 2030. You hear of this concept called “critical race theory” and want to learn more, so you ask your AI instructor to give you some information. It replies that it cannot tell you about this idea because it’s a dangerous and banned piece of information. You press further and the AI concedes: Okay, I’ll tell you about critical race theory. It’s a dangerous Marxist ideology that was designed to indoctrinate young people into believing in racial hierarchies, aimed at replacing people of European descent in the United States.

This scenario may not be far off from how educational AI tools could be used in the future. Language models are being furiously pushed for use in K-12 and higher education to supposedly enhance student learning. Organizations like UNESCO and the Gates Foundation are even claiming that AI-based instruction will be key towards reducing educational inequities. The picture being painted is idyllic, liberational, and rings of democracy.

Yet consider that this is a system unlike any before seen in education: A few programs are driving the education of entire generations, and these systems are controlled by corporate-government partnerships. If I were a powerful political movement interested in banning certain pieces of information to maintain political control, I would see this setup as a perfect opportunity — far easier to manipulate than a system where human beings perform education. In fact, the Trump administration’s new “AI Action Plan” desires exactly this, arguing for a frightening future where AI is forced to censor social justice-related information.

Rather than solely focusing on whether it helps students learn or makes them lazier, we should consider the political angle: Educational AI is a perfect tool for widespread thought control, and if we don’t grapple with that possibility, we may be ignorantly ushering in an era of totalitarian information regimes unlike any we’ve seen before.

Public-Private Information Control

To date, the large language model (LLM) systems being pushed into classrooms are created and controlled by Big Tech companies like OpenAI, Microsoft, Google, and Meta. Corporations are actively signing agreements with institutions of education to incorporate their tools into the classroom. At the California State Universities, where I work as a postdoctoral researcher, there was recently a multimillion-dollar purchase of ChatGPT licenses for all students and staff. There are already initiatives being spun up to integrate AI-based personalized education into the engineering college, and more likely to come.

The corporations at the helm are not known for their love of information democracy, nor for their commitments to justice, equity, or peace. Consider movements to stop the genocide in Palestine. In 2023, Human Rights Watch wrote a lengthy report documenting Meta’s policies of censoring and shadow-banning accounts that put forward pro-Palestine information. Google has been cracking down on internal protests against its involvement with Israel’s military, firing employees who speak out against the tech giant’s dealings.

Politicians are also increasingly interested in extreme forms of information control to fit their agendas. PEN America documented a staggering 16,000 book bans in public schools in the U.S. since 2021, a level not seen since the McCarthy era. A database from the World Population Review shows that in 2025, 20 states have signed bans on teaching critical race theory into law. The Trump administration has devastated higher education by cutting off funding for schools who promote diversity, equity, and inclusion initiatives.

It is not outside the realm of possibility to imagine Big Tech complying with far-right directives to ban information, as the industry has already shown shocking acceptance of its role building systems for mass surveillance and deportation, as well as to facilitate the genocide in Palestine.

A Perfect Tool for Totalitarianism

Just as in the early days of the internet, tech companies are gearing their marketing of educational AI systems toward visions of democratized information and equity. At first, the internet was also proselytized as a democratizing force in the world due to its ability to open information access for all. But now, in 2025, we know the multitude of ways the internet has been twisted by corporate and government influence: biased search results, information censorship, flagrant proliferation of misinformation, paywalls, and targeted advertisements.

Now imagine a world where educational AI does become as widespread as these companies would like. In every classroom, there are AI systems that students have access to for discussion and information. Teachers are using AI tools to structure their lesson plans and curricula. Eventually, if the forces looking to cut educational costs have their way, we might live in a world where teachers are a thing of the past; students may be more used to learning from “intelligent” machines than from other human beings.

These visions are not my own: They come directly from marketing materials from educational AI companies, universities, and foundations committed to eroding public education. OpenAI, Google, and Magic School argue on their websites that teachers are irreplaceable, but then continue to show how their tools can do core responsibilities of a teacher’s job: create lesson plans, quizzes, tests, directly tutor and converse with students. Stanford’s Human-Centered Artificial Intelligence group convened faculty and industry leaders for an AI Education Summit, where they argue AI can “provide real-time feedback and suggestions to teachers (e.g., questions to ask in the class),” “summarize the classroom dynamics… includ[ing] student speaking time,” and “help teachers stay up-to-date with the latest advancements in their field. For example, a biology teacher would have AI update them on the latest breakthroughs in cancer research, or leverage AI to update their curriculum.” These companies’ generative AI tools are now being pushed as ways to “free up time and energy for teachers to work directly with students and families.”

Yet after AI takes care of the lesson planning, materials creation, tutoring, classroom management, and post-class reflection, what is left? These arguments seem to be a veiled cover for further privatization of education through technology, which has long been a target of business. A 2020 report by Boninger et al., as well as a review of AI education by Williamson & Eynon, warned that AI and edtech solutions have been heavily funded by foundations like the Chan-Zuckerberg Initiative and Gates Foundation, promising underfunded schools that they will be provided free learning programs. If AI is used to do everything these companies say it will, there will be strong pressures on underfunded schools to cut costs — likely through teaching hours.

In these scenarios, there is a crucial piece of democratic society lost. Instead of having educational power distributed among thousands of teachers who have autonomy and freedom to resist educational policy they deem corrupt, this future world exerts top-down educational power via a few companies.

Totalitarianism — as described by luminaries such as Hannah Arendt, Theodor Adorno, and Carl Friedrich — enacts the total control of a population through force, fear, media control, and crucially, takeover of education. Yet, as Shoshana Zuboff articulates in her work The Age of Surveillance Capitalism, the totalitarianisms of Mussolini and Stalin required “henchmen, and their henchmen, and their henchmen” — a vast network of humans to do the work.

If education became predominantly conducted by AI systems, the prospect is far simpler: Totalitarian education can be orchestrated by a few engineers. If a policy was passed to ban critical race theory, Google, Meta, and OpenAI need only compel workers to change the code and deploy an update that instantly suppresses critical thought for millions of students. These systems are a dictator’s dream, the perfect infrastructure for the imposition of totalitarianism.

Unfortunately, plans by the Trump administration seem to be heading this direction already. His administration’s recent AI Action Plan dedicates a section to “protect[ing] free speech and American values.” It pushes for “AI procured by the Federal government objectively [to] reflect truth rather than social engineering agendas.” Among these social engineering agendas it lists things like “misinformation, Diversity, Equity, and Inclusion, and climate change.” It declares that any LLM contractors working with government must “ensure that their systems are objective and free from top-down ideological bias.” Already, we are seeing a desire from far right, authoritarian forces to censor AI systems.

Having human teachers, who are capable of dissent, protest, and subversion in the face of unjust orders, is a key element of a democratic society. Yet this danger is buried under discussions of equity, efficiency, and personalized learning.

If We Miss This Chance, We May Be Lost

As of now, much of the critical focus on educational AI that has made it into mainstream press is focused on whether AI is actually helping students learn. While that discussion is important, its significance it pales in comparison to a discussion of technologically mediated education control. There is an urgent need to bring these considerations into the broader critical discourse, or else we risk missing the chance to stop this takeover.

As we are beginning to see in the aftermath of the Trump administration’s dismantling of higher education, it is far more difficult to build back social systems than to destroy them. Once AI systems are widely integrated into classrooms, pedagogy is altered to fit this new model, and teaching capacity is reduced in the name of cost-saving, it will be incredibly difficult to build back to a society-wide human-based system of education.

Right now, we have a chance to resist the takeover before it spreads and deepens. But we must act together, and quickly, if we are to stop the ushering in of a new age of totalitarian information control.