Google Classroom's Real Class

First, in Brief

On 3 November 2026 Google announced that it was changing its artificial-intelligence policy for student accounts in schools. The New York Times broke the story; Futurism, TechCrunch, Engadget, Quartz and the Philadelphia Inquirer, which republished the NYT text, confirmed it the same day. The substance is this:

• The artificial-intelligence tools are off by default. Gemini, NotebookLM, Gemini inside Classroom, and Google Search will be closed on primary and secondary (K-12) student accounts unless a school administrator explicitly turns them on. The change comes in "over the coming months", from the start of 2027.
• But Search is counted as part of the artificial intelligence. In the words of Chris Phillips, Google's vice president for education: "Search and AI are deeply integrated." In a school that restricts artificial intelligence entirely, the student will not be able to reach Google Search either, and the screen will say "your school has turned this off".
• The menu of options has two items. In schools that allow artificial intelligence, the student will use either search with AI summaries on, or search through Guided Learning, the student version of Gemini. There is no option called "classic search, without summaries".
• Teachers will be able to see the chat transcripts. In the school version, teachers will be able to inspect the full transcript of a student's exchanges with the chatbot.
• There is a revolt in the background. In September the two largest school districts in the United States, New York and Los Angeles, suspended students' use of generative artificial intelligence; last week the state of New York passed a law banning some chatbot features for minors.

Google presents this as "giving school administrators control and confidence". Futurism calls the same step "holding schools hostage". In this piece we will explain why both descriptions can be true at once: how the monopoly builds the language of "freedom of choice".

Our thesis, in one sentence: What is given to schools is not control. It is the right to choose from a menu the monopoly has written. Binding search to artificial intelligence is a classic tying arrangement, and it turns the school into a captive customer of the artificial-intelligence market.

---

What Happened? A Chronology

It is misleading to read the event as a single announcement. This is the last move in a contest of about four months.

Date	What happened
19 July 2026	The children's media organisation Common Sense Media rates Google's AI search an "unacceptable risk" for young users. It reports that search does the homework, repeats misinformation in an authoritative tone, and opens to children the Gemini models for which Google sets an age limit in other products. The most important finding: there is no setting that turns the AI off and leaves search on. It is all or nothing. (The Next Web)
August 2026	In schools that have allowed Gemini, Google turns the artificial-intelligence features in Classroom on by default. School administrators react.
September 2026	New York City suspends generative artificial intelligence for students from pre-school through 8th grade for one year (about 600,000 students). Los Angeles also stops students' use of generative artificial intelligence on school devices "until a new decision". (EdWeek, Stateline) Los Angeles administrators hope students will be able to search with AI summaries turned off. Google does not allow it.
October 2026 (final week)	Governor Hochul of New York signs a law banning artificial-intelligence companies from offering some chatbot features to minors.
3 November 2026	Google announces the new policy: artificial intelligence off by default, but if you keep it off, search is off too. The same day Guided Learning opens worldwide inside Classroom; the "step-by-step help" to be added to search begins trials this month in selected US schools.

The table shows this: Google imposed first (in August), and when the reaction grew it appeared to step back (in November), but it stepped back by making search itself the bargaining chip. The default setting changed. The tie did not.

---

Comparing the Sources: One Story, Two Frames

The accuracy of the news is not in dispute. The details overlap across five outlets. What diverges is the frame.

The frame of Google, and of part of the technology press	The critical frame
"Artificial intelligence is now off by default; schools decide as they wish." (TechCrunch's headline: Google will let schools turn AI off)	"It is threatening to cut off access to search; it is holding schools hostage." (Futurism)
Phillips: the aim is to give school leaders "the control and the confidence to use AI as they see fit".	Engadget: the move is an effort to increase the use of artificial intelligence; search is now tied to AI products.
Guided Learning does not answer, it teaches; it supports "productive struggle".	An NYT reporter showed that the consumer version of Guided Learning could write a detailed AP history assignment.
Teachers can see the chat transcripts and so prevent cheating.	This is the recording of the student's process of thinking: a new layer of surveillance.
Schools can use another search engine if they wish.	Chromebooks and Google's education apps are the most widespread systems in US schools. Exit exists in theory. In practice it is expensive.

What matters here is that the left does not fall for the easy line "Futurism is right, TechCrunch is wrong". What Google says is technically true: artificial intelligence is closing by default. But the real question is not the default setting. It is who draws the options, and for whose interest.

---

Tying: The Monopoly's Oldest Trick

In economics this is called tying: selling a product in which you dominate the market together with another product in which you face competition, or whose market you are trying to build; forcing whoever wants the one to take the other.

The best-known historical example is the United States v. Microsoft case of the late 1990s. Using its Windows monopoly, Microsoft declared Internet Explorer "deeply integrated" into the operating system. The defence was familiar: "The browser and the operating system cannot be separated." Today Phillips builds the same sentence for search and artificial intelligence.

The real irony is this: Google itself, in the case brought by the US Department of Justice, was found guilty in August 2024 of illegally maintaining a monopoly in the general search market. So "search" is already a monopoly certified in law. That monopoly is now being used as a lever to win Google customers in the artificial-intelligence market.

Why should it be technically impossible to turn the AI summaries off and leave search on? It is not. Google offered search without summaries for years; even today it does not show a summary for certain kinds of query. "Inseparable" is not an engineering fact. It is a business decision. In July, Common Sense Media asked for exactly this: a switch that turns the artificial intelligence off. Four months later Google gave a switch, but that switch turns search off too.

A reminder for readers in Turkey: on 2 November 2026 the Competition Authority imposed an interim measure on Google's AI summaries. The ground was that the labour of news publishers was being devalued by the summaries. The school move is another front of the same product. There, the publisher's labour was being pulled into the summary. Here, it is the student's attention and habit.

---

Why Is the School So Valuable? Raising the Customer in Line

To understand why a company bargains so hard over children's search box, three layers have to be looked at.

First layer: habit is a property

In Capital, Marx describes how the machine adapts the worker to itself, how the worker is trained to the machine's rhythm. Today's digital monopolies start that training before the workplace, in the school. A child who at ten matches the word "knowledge" with an AI summary has, by twenty-five, already largely learned which chatbot to subscribe to. Here the school is seen, before it is an educational institution, as a nursery for the future customer base.

The ecosystem built with Chromebooks and Classroom does exactly this: student, teacher, assignment, grade, email, document, all inside the same company. This is the school form of the digital enclosure we described in A Common Product Is Not Common Without Common Ownership: a common need (access to knowledge) is closed inside a private property, and then a condition is set for passing through the door.

Second layer: the process of thinking becomes data

The sentence "the teacher can see the chat transcripts" is presented as an innocent pedagogical tool. But who sees the transcript before the teacher? The owner of the server. Which question the student got stuck on, how many times they asked again, where they gave up: this is now measurable data in a company's hands. We took up the systems that measure the second a worker stands still on the factory floor in The Card Around Our Necks, and the artificial intelligence that becomes the boss's eye in When AI Becomes the Boss's Eye. What happens in the school is that eye being fitted in childhood.

Third layer: cognitive surrender is a market strategy

In Whom Is the University For? we discussed the concept of "cognitive surrender": the handing of the labour of thinking over to the machine, the training of the person to the ready answer. The studies cited in the news also show that chatbots can weaken critical thinking and can feed emotional dependence in adolescents. This is why Los Angeles administrators asked for search without summaries: the summary deters the student from looking at more than one source.

The classic form of the search engine was a list: ten links, ten different voices, the labour of comparison. The summary compresses that labour into a single authoritative paragraph. This narrows the field of the capacity to compare and to doubt, which is the core of education. For a monopoly this is not a defect. It is an advantage: the user who does not compare sources is the user who does not question the source.

---

The Class of the Word "Control"

The key word in Google's statement is "control". In whose hands is this control? Let us ask the three questions of Knowledge Commons.

In whose hands? Google defines the options. A school administrator can only tick one of the two or three paths Google has drawn. There is no right to change the menu itself.

Under whose control? Neither teachers, nor parents, nor students can control how these products are designed. Which sources an AI summary is compiled from, which model is suitable for which age group, how long chat transcripts are kept: these are company decisions. There is no public mechanism of control. Where there is one, it is a law running after the company (the New York law, for example).

To whose benefit? If the school says "yes" to artificial intelligence, Google's AI-use statistics rise. That serves to sell investors the enormous investments we discussed in Bubble or Debt? Meta's 145 Billion Dollars, with the line "look, usage is growing". If it says "no", the school is left without search, and has to spend money, time and technical labour on another solution. In both cases the cost is at the school, and the gain is at the monopoly.

Liberal language calls this "consumer choice". But choice is meaningful only between parties of equal power. The relation between a school district and a company whose market value is measured in the trillions is not a bargain. It is a polite form of imposition. What Marx said of the wage contract holds here too: in law, two free parties; in fact, one of them without an option.

---

We Are Not Against Artificial Intelligence. We Are Against the Monopoly

Let us head off a misunderstanding at the start. Knowledge Commons does not argue that artificial intelligence has no place in education. In many of our pieces, beginning with the Socialist AI Manifesto, we have said the opposite: these tools, produced from humanity's common accumulation, can enrich education under social ownership and democratic control. The idea of an assistant that teaches step by step is not bad in itself.

The problem is not in the tool. It is in three places:

1. In the tying: holding a common need (search) hostage to the product one wants to sell (artificial intelligence).
2. In the ownership: the infrastructure of education taking shape on a single private company's servers, according to its commercial calendar.
3. In the control: pedagogical decisions being taken not in teachers' and parents' councils, but in the company's product team.

Those who belittle the suspension decisions of New York and Los Angeles as "hostility to technology" are wrong. These decisions are a public institution trying to use the right to say "on what condition I will accept this product". Google's answer showed how narrow that right is.

---

What Does It Mean for Turkey?

The news is centred on the United States, but it concerns us directly, for three reasons.

First, the same products are used here too. Google Workspace for Education and Classroom are widespread in Turkey, especially in private schools, colleges and universities. Gemini being offered free to students was also announced intensively in Turkey this summer. It is unclear how the settings that come into force in the United States at the start of 2027 will apply to school accounts in Turkey. Google's decision is a decision about a global product.

Second, the digital infrastructure of public education faces the same question. The Ministry of National Education has its own platforms, but when it comes to search, video, documents and email, where the student and the teacher actually go is plain. Unless a public digital education infrastructure is built, when the moment of decision comes, schools in Turkey will sit at the table with even fewer chips than Los Angeles has.

Third, the child-protection discourse exists here too, but in the wrong place. Age verification on social media, brought in in Turkey in the name of "protecting children", builds an infrastructure of identity and surveillance, as we described in How Will Age Verification on Social Media Work, Technically? and The Child-Protection Excuse. Yet there is no voice on the monopolistic education technology that actually touches children's cognitive development. The state's protective reflex turns toward supervising the user, not toward supervising the company.

---

Concrete Tasks

For teachers and education unions:

• Find out which Google account settings are on in your school. Ask whose hands the administrator panel is in. This decision should not be one an IT officer makes alone. It should be on the agenda of the teachers' council.
• Education unions should carry the demand for a pedagogical-impact and data report on the digital platforms used in schools onto the agenda of collective agreements and of the councils.
• Before accepting the opening of chat transcripts to the teacher as "preventing cheating", ask how many days the data is kept, and whether the company has access to it.

For parents:

• In parent-teacher associations, open a discussion of from what age, and on what conditions, artificial-intelligence tools will be turned on. Parent organisations such as Schools Beyond Screens in Los Angeles showed how this discussion can be started from below.
• Show your child that a summary is not a single source, and why it matters to look up the same question in three different sources.

For computing workers:

• Produce practical guides, in Turkish, on free and open-source alternatives for schools: learning-management systems that run on one's own server, open search indexes, and public web-index initiatives such as OpenWebSearch.eu in Europe.
• Prepare "school digital infrastructure" proposals for municipalities and public institutions: a model that does not depend on a single company, and whose data is kept inside the country and under public control.

For all of us:

• Follow the Competition Authority's interim measure on AI summaries, and build public pressure so that the education dimension becomes part of this investigation too.
• Raise the demands that define "search" not as a company's product but as a public infrastructure of knowledge. As the library is public, so the catalogue of the digital age should be public.

---

Dear young comrades,

When you were told at school to "do research", the research itself was in fact a series of labours: finding sources, reading, comparing, noticing that one contradicts another, and then building your own sentence. This labour is tiring. But it is what thinking is.

What Google told the schools this week was, roughly: "Either you will hand that labour over to our machine, or we will not open the door of research at all." It sounds like a technical change of setting. In fact it is a decision of ownership, about on whose conditions access to knowledge will take place.

When monopolies make their own products compulsory, they always use the same word: "integrated". We propose another word: the commons. Knowledge is the common labour of humanity over thousands of years. The road to that labour cannot be a company's bargaining chip. Just as the door of the library is not tied to anyone's purchasing habit, neither should the search box be.

The machine is new, and the chain is old. Whoever recognises the chain finds the way to take the machine into their own hands too.

---