About 30 people stood outside Google office in the West Loop in Chicago on November 11, chanting against Gemini. They were parents, students, educators and school board members, and their complaint was not really about a chatbot. It was about the search box their children use for homework, and a message from Google that the box now comes with AI built in or it does not come at all.
Chicago Public Schools had already blocked ChatGPT, Claude and the Gemini app for students. Days earlier, Google had said schools could turn its AI tools off, but they could not keep Google Search without the AI inside it. That is the choice now facing the three largest public school systems in the United States, and all three are talking in public about walking away from Search.
The Bigger Story
Start with what Google actually announced, because the headline version flattens it. In a November 3 blog post, Chris Phillips, Google vice president and general manager of education, wrote that access to Gemini, Gemini Notebook, Gemini in Classroom and Google Search “will be entirely off for K-12 student accounts until school administrators explicitly turn them on.” Google said it would share more detail in the coming months before the rollout.
That sounds like a win for schools that wanted a pause button. Many of them had asked for one after Google turned Gemini features on inside Classroom in August. The catch arrived in interviews the same day. Phillips told reporters that Search and AI are deeply integrated, and that a school that wants no AI at all will not have Search on student accounts. Students at such a school will see a message saying their school has not made it available.
Google clarified the switches later that day. A district can turn on Search and leave the separate Gemini apps off. What it cannot do is turn on a plain version of Search with AI Overviews and AI Mode removed. In early 2027, reporting on the plan describes two forms of Search for schools that accept it: the standard form with AI Overviews and chat style answers, or a form shaped by Guided Learning, which is meant to walk a student through a problem step by step instead of handing over a finished answer. A pilot of Guided Learning in Search was due to start in selected US schools later in November.
So the argument is not about whether a switch exists. It is about what the switch controls. Schools wanted a switch for AI. Google gave them a switch for each product, and then told them one of those products is now an AI product by design.

Context
Search in a classroom was never just a convenience. For a generation of students, learning to search meant learning to choose a source, open it, compare it with another source, and decide what to trust. Teachers built lessons around that friction on purpose. An AI Overview changes the order of the work. The summary arrives first, in a confident voice, before a student has opened anything at all.
That worry is not only theoretical. Common Sense Media Youth AI Safety Institute rated AI Overviews and AI Mode as an unacceptable risk for kids and teens in an assessment published this year. Its researchers ran more than 2600 test interactions from May 19 to July 1, 2026, using accounts set up as children, and they reported failures on seven of eight AI principles in their framework. In their academic integrity tests, AI Mode completed all 180 homework assignments they gave it. The institute also reported that about three in four US teens and tweens now engage with AI answers that appear in search results, and that a large share of citations in AI answers pointed to user generated material such as forums and social posts.
Google disputes parts of that assessment. The company told reporters the test questions were too narrow and did not reflect how people normally search, and it pointed to protections already built into Search. Both things belong in the record. The tests were designed tests, not a census of real student searches, and their results varied when questions were repeated. They still give districts a concrete reason to ask for a plain search option, especially for younger children, and districts did not invent that concern after November 3. New York City had already placed a one year pause on student facing generative AI from preschool through eighth grade this school year, with high school use limited to approved and supervised settings. Los Angeles had placed a moratorium on generative AI on student devices. Chicago had blocked the best known chatbots while it weighed contracts and privacy terms.
There is a second piece of context, and it is commercial. Google is the dominant supplier of school technology in the US, through Chromebooks, Classroom, Drive, Docs and Gmail. Changing a default search engine on managed devices is, as Los Angeles tech chief Jack Kelanic put it, something a district can do essentially overnight. Leaving the wider Google system is not overnight at all. Rory Mir of the Electronic Frontier Foundation has pointed out that schools sign long contracts and invest heavily in staff training, so a full break carries real money and labour costs. That is why this fight is narrow and revealing at the same time. Nobody in the big three is talking about ripping out Chromebooks this term. They are talking about the one door students walk through most often.
The same week brought a harder reminder of what happens when an AI judgement reaches a student before an adult can check it, in a very different setting. In Students Sue After Doritos Flagged As Gun, a cancelled alert still led to an armed response. The tools differ and the stakes differ, but the pattern rhymes: once an AI output is placed at the front of a school process, the human correction has to travel faster than the output did.
Different Perspectives
New York City has been the bluntest. Schools Chancellor Kamar H. Samuels said in a statement reported by The New York Times on November 18: “If Google doesn’t reconsider their position, we will explore other search options.” He added: “We will not accept the false binary that Google has presented us with.” Mayor Zohran Mamdani accused Google of using monopoly power to push AI into public schools. City Council leaders wrote to Google chief executive Sundar Pichai on November 13 and asked for written answers by November 19. The district also turned down a place in the Guided Learning pilot, according to reporting on its response.
Los Angeles Unified has moved furthest in practice. The district has already turned off AI Overviews for students and says it is reviewing whether Google Search belongs in classrooms at all. Kelanic told a school board AI committee meeting, as reported from the Times account, “We have choices,” and that the district can change the default search engine literally overnight.
Chicago is trying a third route first. A district spokesperson told Chalkbeat that Chicago Public Schools was “actively exploring all options to independently suppress” the AI Overview at the top of searches, and that if it cannot offer students Search without that feature, it will “look to other search solutions.” Outside the office that day, Cassie Creswell, executive director of Illinois Families for Public Schools, which led the rally, argued that if AI is so embedded in Search, schools may need to step back and reassess whether to use Google at all. A student quoted by Chalkbeat put the classroom version plainly, saying students are owed lessons and feedback from the people the district hired, not from a bot.

Google side deserves a fair hearing too, because its case is not nothing. The company says the new defaults give schools more choice, not less, because nothing turns on until an administrator chooses it. It says Guided Learning is built from learning science and is designed to ask questions and offer explanations rather than finish the work. In Classroom, students cannot simply switch that mode off inside an activity and open a freer conversation to get around it, according to Google research scientist Kevin McKee, who acknowledged that AI models were built to complete tasks for users while schools need something different.
Google also shared preliminary research for Guided Learning. In a randomised trial run with the University of Tokyo and described at a press briefing, students using Guided Learning improved test scores by 8.1 percentage points, compared with 5.3 points for students using conventional online resources. That is a real signal, and it is also a limited one. The study covered half a day of independent study, the full technical report had not been published at the time of the briefing, and a gain right after using a tool is not the same as knowledge that lasts when the tool is gone. Google said longer studies are on its roadmap. Districts are being asked to decide before those studies arrive.
The alternatives are not hypothetical. Both Microsoft Bing and DuckDuckGo let schools turn off AI generated answers, according to reporting on the districts search for options. DuckDuckGo chief executive Gabriel Weinberg publicly encouraged schools to consider switching. That single setting is the whole dispute in miniature. Rivals treat AI answers as a feature with a switch. Google treats them as part of Search itself.
Why It Matters
This is a culture story disguised as an admin setting, because it asks who gets to decide how a child first meets information. For years, schools complained about specific apps and then negotiated access, age limits and privacy terms for each one. Search sat underneath that argument, treated as neutral plumbing. Google announcement ends that pretence, in a useful way. If Search is an AI product, then Search needs the same debate every other AI product gets, about accuracy, about homework, about mental health warnings, and about whether a summary should speak before sources do.
It also shows the limit of opt in language when a platform is this large. A district can refuse, in the formal sense. It can also discover that refusing means losing the research tool its teachers, lessons and devices were built around, in the middle of a school year, while a rival search engine has to be tested, approved and explained to families. That is a choice, but it is not a free one, and former Federal Trade Commission chair Lina Khan and Senator Edward Markey are among the public officials who have argued that the structure itself is the problem. Markey wrote to Pichai on November 9 asking Google to reverse its position, citing the Common Sense Media assessment.
For students, the practical stakes are smaller and more daily than the policy fight suggests. Will the first thing they see be a list they have to judge, or an answer they have to doubt? Doubt is a skill, and older students can learn to apply it to an AI summary just as they learned to apply it to a Wikipedia page or a sponsored result. Younger students are still learning what a source is. A system that gives every age group the same front door, with the same summary on top, asks teachers to do the separating work that the product says cannot be done.
Closing Reflection
No district in the big three has switched yet. That matters, and this article should not pretend a threat is a departure. What has changed is that the threat is now credible enough to print, from officials who run schools for well over a million children combined, about a product schools once treated as settled infrastructure.
Google may yet offer a plain Search option, or Guided Learning may prove good enough that districts accept the bundle on its merits. Either outcome would be easier to trust if it arrived as a genuine choice. Right now, the most telling sentence in the whole story belongs to Google itself: Search and AI are deeply integrated. Schools heard that sentence clearly. Their reply is just as clear, and it is not anti technology. It is a request to be allowed to separate learning from answering, at least until a child is old enough to know the difference.
Reporting draws on Google November 3, 2026 announcement by Chris Phillips and its same day clarification, The New York Times reporting on November 3 and November 18, 2026 as relayed and quoted by multiple outlets, Chalkbeat Chicago reporting on the November 11 protest and on Chicago Public Schools position, Forbes reporting from the November 3 Google briefing, and Common Sense Media Youth AI Safety Institute risk assessment of Google Search AI Overview and AI Mode, checked November 2026. Company claims and preliminary research findings are described as such.
Sources
- Google, Designing responsible AI for learning, Chris Phillips, November 3, 2026
- The New York Times, reporting on Google school controls and on New York City, Los Angeles and Chicago responses, November 3 and November 18, 2026
- Chalkbeat Chicago, protest outside Chicago Google office and Chicago Public Schools statement, November 12, 2026
- Forbes, Dan Fitzpatrick, Schools Will Lose Google Search If They Say No To Google AI, November 3, 2026
- Common Sense Media Youth AI Safety Institute, Google Search AI Overview and AI Mode Risk Assessment, 2026
- Hero and body illustrations: Dirty Static
END OF TRANSMISSION
Stay curious. Question everything.


