There are two sides to it.

Attackers use AI. It can help them write phishing emails that read cleaner and feel more personal, imitate a writing style or a voice, and send more of it, faster. That’s a plausible pathway worth planning for. It isn’t a proven driver of education incidents, and we don’t have data tying AI to the ones in our most recent report.

Your people use AI. Unapproved or loosely governed tools can expose student, employee, financial, or research information when someone types it in without knowing how it’s handled. Nobody has to be careless. It can just be a busy Tuesday.

Neither side is brand new. Credential theft, social engineering, and data exposure are the same problems schools have faced for years. AI changes their speed and believability. It doesn’t replace them.

Schools run on distributed networks and distributed people. Students, teachers, staff, and contractors log in from classrooms, homes, and everywhere in between. Remote administration is routine, IT budgets are tight, and the data is some of the most sensitive there is: kids’ records.

That mix shows up in the numbers. Education accounted for 16% of the incidents Huntress saw in 2025. That figure covers incidents across the organizations Huntress protects, not every incident in education, but it tells you schools are on the list.

The report also points to quieter methods. Attackers abuse trusted tools already on the network, and they log in with stolen credentials. When activity looks like a real user doing normal work, nobody hears an alarm.

The typo test is dead. For years, “look for the spelling mistakes” was decent advice. AI can help attackers polish those mistakes away, so your defenses can’t rest on a message looking sloppy. Start here instead.

Set AI-use and data-handling rules

Approve tools through a review process. Spell out what people can and can’t enter, and make those expectations clear to students, staff, and vendors. If the rule is “use your judgment,” it’s not a rule.

Harden identities

Require multi-factor authentication (MFA), starting with administrator and remote-access accounts. Review permissions and pull access people no longer need. Stolen credentials work best on accounts nobody’s watching.

Teach verification habits

Train staff and students to check unusual requests through a separate, known channel. That means password resets, payments, account changes, and anything urgent. Use AI-written text and voice impersonation as examples, but don’t teach people that every message is a trap. The goal is a pause, not paranoia.

Inventory tools and access

You can’t protect what you can’t see. Track AI services, devices, cloud apps, third parties, and remote-management tools. Give each user only the access they need, and watch for activity that doesn’t fit.

Keep the basics sharp and rehearse the response

Patch systems. Protect endpoints. Monitor identity and cloud activity. Keep backups, and test them. Then run your incident response plan before you need it, because the first time shouldn’t be the real time.

Plan for a small team

Most school IT teams can’t do everything. Pick the highest-impact controls first. Decide who investigates, who responds, and who communicates, including nights and weekends. Attackers don’t wait for the bell.

Treating every AI tool as a threat won’t work. Your teachers will use them anyway, just where you can’t see.

So know which tools are in use. Know what data can go in. Know who’s watching for trouble. Lock down identities, cover your endpoints, and practice the response.

Do those things, and AI in school becomes something you manage, not something that happens to you.

Running security for a school or district? See how Huntress works with education organizations to spot threats sooner and respond faster, so your team isn’t carrying it alone.

Want to see what attackers are actually doing in education? Our report breaks down the attack patterns Huntress saw and the practical ways to build resilience, even with a small security team.