Two weeks ago, in a quiet corner of Google’s support forums, an IT administrator at an elementary-middle school did something that many educators have wanted to do for years: they asked Google to give schools a way to block AI Overviews—the chatbot-generated summaries that now appear at the top of search results. The administrator explained, with a mixture of frustration and professional clarity, that these AI answers were undermining the very skills schools are supposed to teach. “This seems counterproductive to what we’re trying to teach,” they wrote. “Students need opportunities to navigate search results, examine sources, and draw their own conclusions rather than being handed an AI-generated synthesis before they’ve even started.” It was a plea rooted in a deeply human concern: that children should learn to think, not just to receive. No one from Google responded to that specific post, but the silence did not surprise the administrator. The concern has been bubbling up for years, with teachers, librarians, and technology coordinators watching as AI features crept into every corner of the digital tools students use daily. In the absence of an official “off” switch, educators from at least three other school districts replied to the thread, sharing their own frustration and describing workarounds they were employing or considering to block AI Overviews. Some used browser extensions. Others tried DNS filtering. A few talked about building custom search portals that would bypass Google entirely. Their expertise, once a niche and somewhat lonely pursuit, was about to become urgently relevant to every school district in the country.
That urgency arrived on Tuesday, when Google announced that it would not give schools an option to disable AI Overviews. The decision, delivered with little fanfare and no accompanying explanation, stunned educators, parents, and technology researchers across the United States. To get rid of AI Overviews, schools would have to do something drastic: block access to Google Search altogether. Not just the AI features. Not just the chatbot summaries. The entire search engine. For districts that have spent years teaching students how to navigate the web, evaluate sources, and build research skills, the announcement felt less like a technical policy change and more like a betrayal of the educational mission. “We were flabbergasted,” said Dominique Johnson, a member of an Oregon community group called Lake Oswego Parents for Responsible Technology. “This move makes things even more difficult for school districts that have determined AI is harmful to foundational learning.” Johnson’s words captured a growing sentiment among parents and teachers that technology companies are making decisions about children’s education without consulting the people who actually understand how children learn. Google, for its part, did not respond to requests for comment. But in a public response on X, the company argued that schools would now have “more choice, not less,” because starting next year, Google Search, the Gemini chatbot, and the Gemini Notebook tool would all be off by default for educational accounts. Administrators, the company said, could enable one, two, or all three services depending on what they felt was appropriate for their students.
At first glance, that promise of choice might sound reasonable. Schools could choose to keep traditional search and disable the chatbot. They could choose to enable Gemini for older students. They could tailor their digital toolkit to their own pedagogical needs. But the problem, as educators and parents quickly pointed out, is that Google considers AI Overviews and a companion feature called AI Mode inseparable from the traditional search experience. AI Overviews, launched in 2023, and AI Mode, launched in 2025, bring the functionality of AI chatbots—such as Gemini or ChatGPT—directly into search results. That means when a student types a question into Google, the first thing they see is not a list of websites, not a collection of primary sources, not an opportunity to compare different perspectives. Instead, they see a paragraph generated by artificial intelligence, synthesized from who knows where, presented with the same clean, authoritative formatting that Google has spent two decades teaching users to trust. For an elementary school student researching a science project, or a middle schooler writing a history paper, that AI-generated paragraph can feel like the end of the journey rather than the beginning. The choice Google is offering, critics argue, is not a choice between AI and no AI. It is a choice between an AI-infested search experience and no search experience at all. That is not empowerment; it is an ultimatum.
The educational consequences are already visible in classrooms across the country. Children are using these search-bar summaries to skip real effort, Johnson said. They pull instant math answers without working through the problem. They ask AI to write their essays and then submit the output as their own work. They copy paragraphs from AI Overviews into their notes without ever reading the original sources. To make matters worse, these AI snapshots frequently make mistakes and output false facts, which younger students take as absolute truth. A fifth grader is unlikely to question a confident, well-formatted answer from Google; they are even less likely to click through to a source and verify. The result is a generation of students who are increasingly dependent on AI for the most basic cognitive tasks, and increasingly incapable of the kind of slow, deliberate, skeptical thinking that education is supposed to cultivate. The IT administrator’s original post struck exactly this chord. Search skills are not obsolete. Learning to read a search results page, to assess a domain name, to distinguish a blog from a peer-reviewed article, to follow a citation back to its origin—these are foundational abilities that prepare students for college, for careers, and for democratic citizenship. Handing them an AI-generated synthesis before they have even started undermines all of that. And while some supporters of AI in education argue that the technology is simply a new tool, educators reply that tools must be used deliberately, not imposed by default. A hammer does not decide to replace a saw; a teacher chooses the right instrument for the task. AI Overviews, by contrast, are always there, always first, always ready to give the answer, regardless of whether the student needs the answer or needs the struggle.
The impact of AI Overviews, however, reaches far beyond the classroom. The AI-generated responses are tempting to read and are often just good enough to trust, and that is causing headaches across the broader information ecosystem. Wikipedia, news websites, educational publishers, and countless other online resources have reported significant losses in human visitors as people increasingly rely on AI summaries instead of clicking through to actual pages. When a student reads an AI Overview, they are not visiting the websites that produced the underlying knowledge. They are not reading the original reporting, the scholarly article, or the primary source document. They are not contributing to the ad-supported revenue that keeps many of those websites alive. Over time, this creates a dangerous feedback loop: fewer visitors means less funding for quality content, which means a thinner and poorer internet, which means AI models have less good data to learn from in the future. Even more troubling is the fact that Google uses the same bots to crawl the web for both its search engine and its AI products. This makes it impossible for publishers to opt out of having their content used for AI-generated answers without also disappearing from Google Search entirely. A newspaper cannot say, “You may index my articles in search results, but do not use them to write AI summaries.” There is no middle ground. This is not just a problem for publishers; it is a problem for everyone who cares about a healthy information commons. When the people who produce knowledge cannot control how it is used, and when the systems that distribute knowledge are optimized to prevent users from ever reaching the original sources, the entire ecosystem begins to collapse.
Ultimately, the debate over AI Overviews in schools is not just about technology. It is about values. It is about whether we believe that children should be taught to think or taught to consume. It is about whether the convenience of an instant answer is worth the cost of a generation that never learns to ask better questions. School districts across the United States are now faced with an impossible choice: allow students to use Google Search, knowing that AI Overviews will be lurking at the top of every result, or block search altogether, depriving students of one of the most important digital literacy tools of the modern age. Neither option is acceptable, and yet Google has offered no third path. There are workarounds, to be sure—custom search engines, alternate browsers, subscriptions to educational databases—but these are partial, imperfect, and often expensive. They do not address the fundamental problem, which is that a powerful company has decided that AI is the future of search, and that schools, parents, and students will simply have to adapt. The parents of Lake Oswego, the IT administrators in school districts across the country, and the teachers who see AI Overviews replacing research skills every single day are not asking for a world without technology. They are asking for a world where technology serves human learning rather than replacing it. They are asking for an off switch, a choice, a way to protect the messy, beautiful, essential process of curiosity and discovery. Until Google is willing to listen, the burden will fall on educators and families to fight for that space—one search at a time.
