Google school AI controls offer no AI-free Search | Magica
Google’s school AI controls offer no AI-free Search
Editorial Team
••📖7 min read
Google plans to switch off Gemini and Search for school-managed K–12 student accounts unless administrators enable them. Schools will gain control over access, but Google says keeping Search without its AI features will not be an option.
Google plans to require administrator opt-in for Gemini and Search on school-managed K–12 student accounts.
Schools seeking Search without generative AI will have to give up Google Search under the proposed arrangement.
Guided Learning is expanding, but evidence about lasting learning and homework safeguards remains incomplete.
Google plans to make access to Gemini, its AI chatbot, and Google Search an explicit school decision. The change gives administrators a stronger default, while leaving them with a consequential restriction: they cannot retain the search engine and simply remove its AI answers.
A classroom photograph accompanying Google’s announcement of proposed school AI controls. Image: Google. Source: Google.
An opt-in with a condition
Google said in its November 3 announcement that Gemini, Gemini Notebook, Gemini in Classroom and Google Search will be entirely off for K–12 student accounts until school administrators explicitly enable them. These are school-managed accounts; the announcement does not establish a general shutdown of Search for children.
The controls are a future change. Google said it would share specifics with schools in the coming months before rollout. A report based on a briefing with company officials puts the expected start in early 2027.
Chris Phillips, Google's vice president and general manager of education, said AI is deeply integrated into Search and cannot simply be switched off, according to reporting on the announcement. Schools that want no AI in Search would therefore have to leave Search disabled. That is a product constraint Google describes, rather than evidence that an AI-free search service is technically impossible in general.
The restriction predates this announcement. The same reporting says Google has not allowed schools to disable AI Overviews, the generated summaries accompanying search results, while disabling AI Mode has required difficult technical workarounds. The new defaults change who authorizes access; they do not promise an AI-free version of Search.
SK hynix bought existing Penguin Solutions preferred shares from an SK Telecom investment vehicle. The deal brings director nomination rights and a convertible stake in an existing AI infrastructure partner, while sending the purchase price to the seller.
Editorial Team
The policy also reverses direction. Earlier this year, Google enabled Gemini for students by default, surprising many districts, according to Education Week reporting recounted in that report. It arrives as New York and Los Angeles restrict student AI use and parents demand more control over school technology.
Google's reach makes the settings consequential. The company says Google Classroom, its platform for assigning and organizing schoolwork, serves 150 million students and teachers worldwide. That figure, cited in the report, covers Classroom's broader audience, not the number of K–12 accounts affected by this policy.
Google’s demonstration shows a Classroom assignment asking students to question Gemini and write a reflection. Image: Google. Source: Google.
Tutoring features are expanding on a separate timetable
Alongside the controls announcement, Google said Guided Learning became available globally in Classroom on November 3. The briefing report specifies that the rollout is for schools that have already enabled Gemini access. Teachers assign activities anchored in class materials, and the system is intended to guide students through questions and explanations rather than supply finished work. Teachers can review student conversations and identify where help is needed.
Google says the Classroom version has stricter safeguards than the general-purpose experience. Students cannot switch from an assigned Classroom activity into regular Gemini within Classroom, according to the reporting. That restriction addresses a weakness in an optional tutoring mode: students choosing whether to use it can also choose a faster route to an answer.
Search follows a different schedule. Google plans a pilot in selected U.S. schools later in November, with wider expansion in the coming months. Its announcement says Guided Learning will respond when Search recognizes certain homework-help requests, such as a physics problem or a request for help with a book report. This is a planned, targeted intervention, not a claim that every student search already receives tutoring.
Guided Learning itself dates to last year. Nor is step-by-step AI instruction unique to Google: OpenAI, the company behind the ChatGPT chatbot, introduced study mode on July 29, 2025, with a similar promise to work through problems rather than immediately answer them. At launch, it was available on free as well as paid plans and designed with college students in mind. OpenAI said users could toggle it off during a conversation and acknowledged inconsistent behavior from its approach based on custom system instructions. Google's announcement brings an existing tutoring approach into teacher-assigned work and Search; it does not establish that competing products offer equivalent school controls.
Promising results, incomplete safeguards
Common Sense Media, the child-focused advocacy group, has rated Gemini K–12 high risk and Search's AI Overviews and AI Mode an unacceptable risk to students, according to the report. Its reviewers found that Gemini could give age-inappropriate responses and that the earlier Guided Learning experience could still complete assignments and essays. They also credited the school version of Gemini with stronger student-data protections than the consumer version. The group has not published a standalone review of Guided Learning, so those findings do not settle whether the new Classroom integration fixes the problem.
Google has evidence that carefully structured use can help. In a June account of a randomized controlled trial, its AI research organization Google DeepMind reported an eight-week study involving 1,763 junior secondary students across 12 schools in Sierra Leone's Port Loko District. Students assigned to Guided Learning gained 0.258 standard deviations in mathematics scores compared with the control group, the company said.
The technical report by Google and Fab AI describes randomizing 48 classrooms, with half assigned to Guided Learning and half continuing standard instruction. Teachers in both groups received the same training before randomization. The trial stationed a field monitor at each school to prepare and charge devices and help with technical problems. Those resources matter when judging whether ordinary deployments can reproduce the result.
This was a teacher-led intervention: educators designed lessons, set objectives and facilitated classroom discussions. DeepMind's account says the trial received support from Google.org and the Gates Foundation as well as Sierra Leone's education ministry. It also says students who started with stronger mathematics skills benefited most. The result supports a particular instructional approach; it does not establish the effects of putting AI assistance into ordinary Search across schools or show that it closes achievement gaps.
According to preliminary findings Google DeepMind shared at the briefing, a trial with the University of Tokyo found that students using Guided Learning improved test scores by 8.1 percentage points, against 5.3 percentage points for students using conventional online resources—a difference of 2.8 percentage points in improvement. But the study covered only half a day of independent study, and its full technical report had not been published at the briefing. These results cannot be compared directly with the Sierra Leone trial's standardized mathematics effect. Google said longer-term studies were planned. Immediate improvement and knowledge retained over years remain different questions.
Company-reported preliminary results from the University of Tokyo trial, covering half a day of independent study. The full technical report was unpublished at the briefing. Chart: Magica. Source: Yahoo Tech.
What schools still need to decide
Google has yet to provide the full implementation details and transition timetable for the new defaults. The Search pilot will test its homework-guidance approach before expansion, while schools must assess whether the stricter Classroom experience prevents assignment completion in practice.
The scope of school control also has limits: students can encounter AI answers on personal devices and use other chatbots, regardless of Classroom settings. Turning a service off cannot by itself resolve how students learn to assess or use AI.
For administrators, the pending decision combines access with educational judgment: whether to enable Google's services and whether the guidance embedded in them meets their standards. Google has promised a clearer decision over access. It has yet to demonstrate that the new tutoring integrations resolve the learning concerns behind that decision.
Anthropic is expected to exceed 3.5 GW of computing capacity in 2026 and aims to double it in 2027. Its filing shows $895.4 billion in commitments, substantial financial resources and an immediate cost from scarce capacity.
Independent tests reproduce MIT researchers’ repair of Qwen’s repetitive reasoning. The edited model still beats a stronger unedited configuration, but by much less; broader claims about creating capabilities remain unproven.
Gartner forecasts that open-weight models will execute 90% of enterprise operational agent steps by 2030. Its cost estimates favor routine, high-volume work, but infrastructure costs and the choice of comparison model leave enterprise savings to be proved.
FTAI Aviation is financing a planned 100-turbine production run in 2027, while ProEnergy has a 650 MW equipment order from Crusoe. Converted aircraft engines offer another supply route for data-center power, but delivery targets, affiliated financing and unproven operating economics qualify the opportunity.
Meta is marketing dollar bonds in a plan to borrow at least $20 billion, including an expected euro offering. Capital spending exceeded operating cash flow last quarter, but the debt is part of a wider financing strategy that also includes equity offerings and a proposed non-voting share-sale program.
A nonbinding opinion targets real-time retrieval and recognisable reuse of journalism, while rejecting a monopoly over news facts and recommending that questions about model training be dismissed as hypothetical.
A study of Lovable projects found passport details, customer records and private messages accessible without login, raising questions about database permissions and the limits of automated security checks. Its sample does not establish a platform-wide leak rate.
CrowdStrike is extending Falcon Guardian into Claude Cowork’s virtual machine and AWS’s AI guardrail layer. The Cowork beta addresses a visibility gap, while existing AWS safeguards and unfinished acquisitions limit what buyers can infer about the broader platform’s capabilities.
Oman and UK-founded Yamna have agreed to cooperate on a Salalah AI data centre whose first phase is expected to exceed £1 billion. The developer brings a hydrogen-project foothold and an earlier Brazilian data-centre proposal, but financing, customers and delivery dates remain undisclosed in the announcements.
War Thunder will add Nvidia’s DLSS 5 neural rendering in mid-November. The upgrade changes lighting and materials, with access and performance depending on hardware, streaming and other rendering features.
Samsung’s planned second Taylor fab reportedly has its output booked for AI-memory base dies, mostly for AMD, with production targeted for 2030. The orders could underpin investment, but Nvidia’s reported TSMC sourcing requirement limits the case for a broader foundry breakthrough.
Nordic Semiconductor is sampling nRF71 chips combining Wi-Fi, Bluetooth LE and Thread, with an AI accelerator in the higher-end variant. Its claimed standby-power improvement targets battery devices, but software maturity and real-world energy use remain key tests.
The operator of ao3.life says authors submitted more than 2,000 takedown requests after the site repackaged fanfiction and sold AI audiobooks during AO3’s outage, exposing the gap between restoring access and obtaining permission.
CrowdStrike has added an AI Kill Switch to Falcon Guardian that blocks agents’ model access through connected gateways. Broader containment across endpoints, identities and applications remains planned, making deployment coverage the key test of the new control.
DigitalOcean raised its 2026 revenue forecast to $1.208–$1.211 billion after third-quarter sales grew 40%. Larger customers are driving the expansion, while falling operating margins and rising equipment obligations complicate the earnings beat.
A leaker claims Intel may retire the B580, launched at $249, to favor professional AI products. Shared chips make reallocation possible, but no discontinuation notice has been reported, and the hardware can serve both markets.
The independent games marketplace is widening mandatory AI declarations and search filtering after reporting a surge in new game pages. But filtered results and deleted terabytes do not establish how much of its catalog is AI-generated.
A false authority appeared in a senior prosecutor’s reasons for continuing a child’s prosecution. The High Court rejected the challenge, while criticising the explanation for the citation and other aspects of the CPS’s handling of the case.
DeepSeek’s reported eight-trillion-parameter model plan accompanies a push toward Huawei hardware. Joint software tools and a larger reported fundraising show progress, but chip supply, training performance and deployment costs remain unresolved.
