Over the past few years, Google’s low-priced laptops have become an AI Trojan horse in America’s schools. Students can use a school-issued Chromebook to automate much of their homework. Fire up the Gemini app, and Google’s chatbot will spit out a full essay about the meaning of The Giver. Type an algebra question into the Google search box, and an “AI Overview” will solve it for you. Ask a follow-up question, and the search box might give way to “AI Mode,” which turns the search engine into a full-on chatbot.
Districts across the country give each student their own device, which typically comes equipped with a Google account and a suite of the company’s software. Over the past year, some of the country’s largest public-school districts, including Los Angeles and New York City, have passed policies limiting students’ screen time and banning generative-AI tools. But they hit an obstacle: Google didn’t offer schools a way to fully opt out of AI.
Yesterday, the tech giant’s attempt to address the problem only made people madder. Google’s stand-alone Gemini chatbot will now be disabled by default for school accounts, the company announced. But Google quietly left out any option to turn off its most popular and widely used generative-AI features: those built into its search engine. In other words, the only way for schools to go AI-free is to cut off students’ access to Google Search, the world’s most popular website.
Tech companies threatening to withdraw products rather than modify them has become part of Silicon Valley’s playbook. (Meta once responded to a Canadian law requiring the company to share revenue with news publishers by banning news articles from its platform altogether.) Online, the move drew outrage, including from congressional Democrats who will soon hold newfound legislative and subpoena powers. “Google is once again prioritizing profits over people,” Senator Richard Blumenthal of Connecticut posted. “Clearly, the Tech Giant is more concerned with forcing Gemini into schools than enabling students to succeed.”
Google contended that the anger was based on a misunderstanding. “Today’s announcements give schools more choice, not less,” the company posted on X, evincing a faith in the power of boldface font to get its message across. That much is true. Google has never offered schools an AI-free version of its search engine (though there is a “Web” tab that limits results to links). What was new in Google’s announcement was that both Gemini and Search will now be turned off by default, and schools will have the option to enable one or both. For schools that enable Search, the company is touting a new “guided learning” mode in its AI features that will prompt students to work through questions one step at a time, rather than simply giving them the answer. That could help address one of the biggest objections to Google’s AI search tools in the classroom, which is that they short-circuit the learning process. (A spokesperson for the company declined to comment beyond pointing me to its public statements.)
[Read: My son’s math homework is essentially just Pokemon]
Yet Google’s response to the blowback failed to address the aspect of its announcement that had upset people in the first place. The company wasn’t taking away any options that had existed before, but it also wasn’t providing the one option schools seemed to want most: a way to let students use Google Search without also enabling the AI bells and whistles.
The unspoken logic behind the company’s stance is that, in some sense, all Search is AI search. AI Overviews and AI Mode launched in 2024 and 2025, respectively, but the company has been using large language models to help rank its search results since well before the arrival of ChatGPT. Google also used AI for years to help power the precursors to AI Overviews, including “featured snippets”—those bolded excerpts from web results that Google surfaced atop its list of links. So when Google says that opting out of AI entails opting out of Search, that’s technically true.
Even so, there’s something obtuse about Google’s refusal to acknowledge that consumers—and schools especially—might reasonably draw a distinction between LLMs that help power search results and LLMs that generate answers themselves. Parents and teachers mostly don’t object to Google using AI to decide which webpages to link to when students ask a question (though there is a case to be made that the company’s opaque ranking process has always deserved more scrutiny than it’s received). They object to turning Google into an instant-answer machine—and an error-prone one, at that. They object to students being conditioned to delegate their critical-thinking and source-evaluation skills to a machine. And they object to infusing the same site that indexes the world’s online information with a chatbot that can function as an ersatz friend or dispense personal advice to impressionable children.
Which brings us to the business logic underlying Google’s stubborn stance. The company built its empire around a search tool that offered the shortest path between a user’s question and the information they were seeking. For years, that mostly meant refining its algorithm to downgrade spammy sites in its results. Then came ChatGPT, which promised to answer just about anything a user could ask, dispensing with the list of links altogether. OpenAI claimed this week that ChatGPT now handles about one-fifth as many search queries as Google. Chatbots are the most significant threat to Google’s core business that the search giant has ever faced.
Its response has been to morph Search into an AI-answer tool in its own right. A few discontented users have dropped Google for niche alternatives such as DuckDuckGo, the privacy-focused search engine that now touts AI-free search as a selling point. Kamyl Bazbaz, a DuckDuckGo spokesperson, told me that usage of the company’s “No-AI” search option is up 30 percent since September. The company is now looking for ways to make it easier for schools that opt out of Google Search to switch their default search engine, he added. But Google seems to be betting that convenience will win out over qualms about generative AI. That bet could easily pay off, given how accustomed students already are to the tools Google has provided them with.
While researching this story, I asked Google for data on how many Chromebooks are in U.S. schools. The site responded with an AI Overview confidently asserting that 93 percent of U.S. school districts plan to purchase Chromebooks. In small print, the overview cited one source for that figure: an article from a website called commandlinux.com. Skeptical, I did what fewer and fewer Google users are doing these days: I clicked on the link. The article, which appeared to be AI-written, offers no methodology or original source for any of the figures it provides.
I took those extra steps because I’m a professional journalist who covers AI and its limitations for a living, and because my own schooling trained me to seek out original sources and evaluate their credibility. Had I grown up with AI search, as today’s kids do, I doubt I would have thought to question the number that Google handed me.
Leave a Reply