Quick Answer: Blocking Google's AI tools doesn't automatically block Google Search, but a poorly scoped block can. Schools that shut off Gemini through a network filter or a broad domain rule can take away search results too, because AI features are woven into the same google.com pages students use every day. The safer route is to disable AI features at the Google Workspace admin level and keep search running.
Schools are reaching for a blunt tool, and the bluntness is the problem. A growing number of districts want to keep generative AI away from students, and the reporting in The New York Times raises the stakes: some of those blocks could cut off the search engine students rely on for research. If you run IT for a school, or you're a teacher wondering why results look different, this is the question to answer before you change anything.
Why a block on AI can reach ordinary search
The core issue is architecture. Google doesn't run "AI" on a separate site that you can filter the way you'd block a social network. Gemini has its own app and domain, but AI Overviews and AI Mode appear on the same search pages students open for a plain query. From a firewall's point of view, the traffic is going to google.com either way.
Domain-level filtering is where most accidental damage happens. A network administrator who adds a broad rule to stop AI content may block the path or host that serves search results, not just the Gemini interface. The result looks like "Google is down" to a student, and it shows up as a helpdesk ticket on Monday morning.
Here's the catch: the exact behavior depends on how Google serves its pages at a given moment, and it changes. Test any rule with a real student account on a real managed device before you roll it out district-wide.
The three layers you can actually control
Most schools have three places to act, and they behave very differently.
- Network and DNS filtering. Tools like content filters or DNS resolvers can block hostnames. This is the bluntest layer and the one most likely to hit search.
- Google Workspace for Education admin console. Admins can turn the Gemini app on or off for specific organizational units, such as students versus staff. This is the layer Google designed for AI access decisions.
- Chromebook and Chrome policies. Managed browser settings can restrict features on school-issued devices, separate from the network.
Think of these as cause and effect. A network block fires on every device, every account, every time, so it touches search whether or not anyone meant it to. An org-unit setting in the admin console changes what a specific group of accounts can use, and it leaves the rest of the web alone. That's the difference between a sledgehammer and a scalpel.
What the compliance frameworks actually require
Schools often assume a block is mandatory. In the US, the picture is more specific than that. The Children's Internet Protection Act (CIPA) requires schools receiving E-rate discounts to filter internet access for minors, covering obscene content, child pornography, and content harmful to minors. CIPA does not require you to block an AI product. It requires filtering that meets its standards, and those standards are compatible with keeping search available.
FERPA and COPPA shape a different question: what data a school shares with a vendor. A school can reasonably decide that a student-facing AI tool needs a privacy review before it's enabled. That's a legitimate reason to restrict Gemini specifically. It's not automatically a reason to remove search.
That said, the requirement to filter is real, and schools often over-correct. The honest move is to document which law requires which control, then pick the narrowest tool that meets it.
Comparing the blocking approaches
| Approach | What it blocks | Effect on Google Search | Main risk |
|---|---|---|---|
| Block Gemini hostname at DNS | Gemini web app only | Usually minimal, but test it | Students use other AI tools instead |
| Workspace admin: turn off Gemini app for student OU | Gemini app for those accounts | None on search | Admin setting changes across product versions |
| Chrome managed policy on Chromebooks | Selected browser features | Depends on policy | Fragmented rules across device fleets |
| Block google.com paths broadly | Search and AI features together | Severe | Students lose basic research access |
The table reflects the general tradeoffs, not a guarantee for every district. Your setup will vary with your filter vendor and your Workspace edition.
Why the common advice is incomplete
The usual advice is "block the AI, keep the search." That's correct in principle but incomplete in practice. Most guides stop at the admin console and assume it covers everything. It doesn't cover network filters that sit upstream, and those can override a careful admin-level setting.
Here's the counter-intuitive part. The more layers you stack, the more likely you are to break search, not less. Each filter is independent, and nobody usually audits the combined effect. A school can turn off Gemini in the console, leave an old DNS rule in place, and end up blocking search for a subset of devices without anyone noticing for weeks.
Most people stop here, and that's a mistake. Audit the full stack, not each layer alone.
A practical sequence for IT teams
Work through this in order. Each step catches a failure the previous one might miss.
- Document the goal. Write down whether you're restricting Gemini, AI Overviews, or all generative AI. These are different targets.
- Set the org-unit control first. Turn off the Gemini app for student accounts in the Google Workspace admin console, and confirm the setting in the policy's audit or reporting view.
- Keep SafeSearch on. Google Workspace allows admins to enforce SafeSearch for managed accounts. This handles explicit content in results without breaking search itself.
- Test on a Chromebook with a student test account. Run ten ordinary research queries, including a question that would normally trigger an AI Overview, and note what appears.
- Review your network filter. Check whether any hostname or path rule touches google.com beyond the Gemini host.
- Set a re-test schedule. Google's interface changes often, so re-run the test after major product updates or filter-vendor pushes.
Here's where it gets interesting. Step 4 is the one most districts skip, and it's the only step that shows you what students actually experience.
The cost of getting it wrong
Consider a middle school with a single network rule added in a hurry to stop AI homework help. Within a day, students report that every Google search returns an error. Teachers assume the network is down. The helpdesk spends hours on a problem that a single rule created. The fix takes minutes once someone finds it, but the lost class time is real.
That scenario is a plausible failure mode, not a documented case from the reporting. It's the kind of mistake that happens when a block is applied at the wrong layer. The lesson holds regardless: a tool that stops a fight at the wrong level can start a much bigger one.
FAQ
Frequently Asked Questions
Can schools block Google AI without blocking Google Search?
Yes, in most setups. Turn off the Gemini app for student accounts in the Google Workspace admin console, and avoid broad network rules on google.com. Test with a student account after any change, because AI features and search results share the same domain.
How do I turn off Gemini for students?
In the Google Workspace admin console, go to the Gemini app settings, select the student organizational unit, and set the app to off. Confirm the change applied by signing in as a test student. Exact menu names change, so check your current console version.
Does blocking Gemini turn off Google Search?
Usually not, if you block it at the Workspace account level. A DNS or firewall rule that blocks the wrong hostname or path can affect search too. Test ordinary queries on a managed device before and after the change.
Is blocking AI in schools required by law?
In the US, CIPA requires schools that receive E-rate discounts to filter internet access for minors. It doesn't name generative AI specifically. FERPA and COPPA affect how student data is shared with vendors, which may justify restricting a specific AI tool.
Why would a school block Google AI in the first place?
The most common reasons are concerns about academic integrity, student privacy, and age-appropriate content. Each concern points to a different control, so define the problem before you pick a tool.
Closing
If your school is considering this, the most useful thing you can do this week is run a ten-query test on a managed student account before and after any change. Keep the restriction at the Workspace admin layer, and treat network rules as a last resort. Pass this to whoever manages your school's filters, and ask them to check whether any rule touches google.com beyond Gemini. Blocking schools block Google AI is a reasonable goal, but only if search survives the process.