School AI Tools and Student Data: What Parents Should Ask

A Virginia high school principal who dropped computer science after one day in college now builds his own apps. Using Google’s AI tools, Derek Cantrell created a system that tracks his students’ unexcused absences, tardies and disciplinary referrals, warns them as they near the limit, and lets them “earn back” good standing. He told Education Week in October 2026 that the key is keeping student data “localized and safe.” That story captures a quiet shift in school AI tools and student data: school leaders aren’t just buying AI software anymore, they’re building it. That can be a real help. It can also skip the checks that purchased software is supposed to go through. Here’s what’s happening and what to ask.

Key Takeaways

• Principals are using AI to build custom tools for attendance, discipline tracking, teacher goals and student information dashboards, according to Education Week reporting from October 2026.
• In an EdWeek Research Center survey of 852 educators (Aug–Oct 2026), the top operational uses of AI among leaders who used it were communications, translation, strategic planning, employee evaluation and class scheduling.
• The Center for Democracy & Technology’s September 2026 toolkit gives school leaders five risk questions: who’s affected, how likely the tool is to fail, the consequences of failure, whether errors can be corrected, and whether risks can be mitigated.
• Home-built tools can bypass procurement review, privacy agreements and accuracy checks that vendor tools are supposed to receive.
• Parents have a right to ask what data a tool uses, who approved it, and how a student can challenge an AI-influenced decision.

What “principals building AI tools” looks like

Generative AI now lets non-programmers build working software by describing what they want. School leaders have noticed. In an October 2026 article, Education Week’s Olina Banerji described several examples:

• Attendance and behavior. Cantrell, a high school principal in Covington, Virginia, built a system that tracks absences, tardies and referrals, warns students approaching limits, and lets them restore standing through lunch detentions or other tasks. Students who stay under thresholds earn points for cafeteria treats. He also built tools for tracking teacher observations and graduation progress.
• Teacher goals. Sterling Seemans, principal of Brandywine High School in Wilmington, Delaware, created a set of AI prompts to help teachers draft yearly goals with measurable objectives.
• Student information. Rachel Jensen, principal of an online public K–12 school in Idaho, uses automated tools that pull student information, including histories of individual needs and accommodations, into trackers for teachers overseeing more than 90 students.

Cantrell’s framing was positive: “people over paperwork.” The tools free time for being “present in the hallways and in the classrooms with kids.” He also offered the essential caution himself: “You don’t want to put student data into things that go out into the public domain.”

These aren’t fringe uses. A nationally representative EdWeek Research Center survey of 852 educators, including 262 school and district leaders, conducted August 26 to October 6, 2026, found that nearly 1 in 3 leaders don’t use AI for operations. Among those who do, the top uses were communications, foreign-language translation, strategic planning, employee evaluation and class scheduling. EdWeek also cites its research center’s data that more than half of districts use AI tools in teacher hiring.

Why home-built school AI tools and student data are a riskier mix

Here’s the engineering issue in plain terms.

When a district buys software, it is supposed to go through a review: legal checks on privacy, a data-sharing agreement, security requirements, sometimes an accessibility audit. It’s imperfect, but it’s a gate.

When a principal builds a tool with a general AI assistant, there may be no gate at all. Three things can go wrong.

Where the data goes. A tool built by pasting student records into a chatbot, or by connecting a spreadsheet to an AI service, may send that data to the AI company. Whether it’s stored or used depends on the account type and terms. School-managed enterprise accounts usually offer more protection than personal ones. California’s model AI policy addresses this directly: “Personal AI accounts should not be used for any purpose involving student data, student work, student records.”

Whether the logic is right. AI-written code and formulas can contain errors that look fine. An attendance tool that miscounts tardies, or applies a threshold to the wrong field, will flag the wrong students. Nobody may notice until a family complains.

Who decides. A tool that tracks absences and assigns consequences is making, or shaping, disciplinary decisions. That’s the category of use that guidance consistently calls high-risk.

The federal rules point the same way. The FTC’s COPPA guidance recommends that “schools or school districts decide whether a particular site’s or service’s information practices are appropriate, rather than delegating that decision to the teacher,” and the same logic applies to principals building tools one school at a time. The US Department of Education’s guidance on online educational services explains when sharing student records with outside services is permitted under FERPA.

The CDT toolkit’s five risk questions

In September 2026, the Center for Democracy & Technology published From Possibility to Practice: A School Leader’s Toolkit for Responsible AI, by Elizabeth Laird, Kristin Woelfel and Hannah Quay-de la Vallee. Unlike most AI guidance, which focuses on teachers and students, it addresses how school leaders oversee AI, including their own use.

Its core is five questions for assessing the risk of any AI use. They work just as well for parents:

CDT risk question	What raises the risk	Applied to a home-built attendance tool
Who will be affected?	Large-scale use, especially on students; outsized impact on protected or vulnerable groups	Every student; students with chronic absence often face other hardships
How likely is the tool to fail or err?	Little information on performance; tools not designed for schools	Built by one person, untested by anyone else
What are the consequences of failure?	High-stakes uses with serious consequences	Wrong warnings or detentions; records that follow a student
Can failures be corrected?	Errors hard to detect or reverse	Only if someone checks the outputs and families can appeal
Can the risks be mitigated?	No human review; no clear safeguards	Human review of every consequence; data kept in district-approved systems

The toolkit gives a simple example of how the same task can be low- or high-risk. Using AI to write an attendance policy “in its entirety with little human review” is high-risk. Using AI to brainstorm scenarios the policy should cover, while a school leader writes it, “would carry much lower risk.”

It also points to existing guidance. Georgia’s Department of Education identifies high-risk uses such as evaluating educators and writing IEP goals. And New York City’s guidelines, according to the toolkit, include uses that are never allowed, “such as using AI for counseling, grading, surveillance, or student discipline.”

The toolkit also documents why trust matters. Among the figures it cites: half of parents and students agree that a teacher who uses AI in class “is not really doing their job,” and half of students say AI in class makes them feel less connected to their teacher.

Questions parents can ask the school

These are polite, specific and hard to wave away. Ask in writing.

”What AI tools does the school use with student information, including ones staff built?”

Many schools keep an inventory of purchased software. Fewer track tools staff build themselves. Asking the question signals that both count.

”What student data goes into each tool, and where is it processed?”

Ask whether the tool runs inside the district’s approved accounts or sends data to an outside AI service, and whether that service stores it or uses it for training. Our explainer on what data schools collect through ed-tech covers the categories to ask about.

”Who approved it, and was it checked for accuracy?”

For anything that affects grades, discipline, attendance records or special education, ask whether someone other than its builder tested it, and how.

”Does a person review every decision the tool influences?”

An attendance warning generated by AI and checked by a human is very different from one sent automatically. California’s model policy gives guardians the right to request “timely human review of any AI-generated outputs that inform decisions about their child,” a useful standard to cite anywhere.

”How can my child challenge a decision?”

There should be a clear way to dispute an AI-influenced record. Our guide to states’ rules on AI notification and opt-outs explains what some states require.

What not to do: assume home-built means bad

Many of these tools solve real problems, and principals like Cantrell are thoughtful about privacy. The point of asking isn’t to shut them down. It’s to make sure they get the same basic checks a vendor’s product should, and that a human stays accountable for decisions about your child. For the vendor side of the same question, see our look at the Microsoft-AFT ban on student data AI training.

What to Watch For Over the Next 3 Months

• Week 4: Check what you’ve received from school this term: attendance alerts, behavior notices, progress reports. Ask whether any were generated or shaped by AI.
• Month 2 red flags: An attendance or discipline notice that doesn’t match what you know, a vague answer to “what tool produced this?”, or no route to have a person review it.
• Month 3 self-check: Do you know which AI tools touch your child’s records? Does the school have a written process for approving tools staff build? Has anyone shown you how to dispute an AI-influenced decision?

Frequently Asked Questions

Are principals really building their own AI tools?

Yes. Education Week reported in October 2026 on principals using AI to build attendance and discipline trackers, teacher goal-setting prompts, and student information dashboards. Generative AI lets people without programming experience create working apps by describing what they want, which has made home-built tools much more common.

Is it legal for a school to put my child’s data into an AI tool?

It depends on the tool, the agreement with its provider, and your state’s laws. FERPA permits sharing education records with outside services under certain conditions, and COPPA lets schools consent for parents only for educational purposes. Personal AI accounts used with student data raise the biggest concerns. Ask your school which conditions apply.

What is the CDT school leader toolkit?

It’s a September 2026 guide from the Center for Democracy & Technology called “From Possibility to Practice,” which helps principals oversee AI use responsibly. Its five risk questions ask who’s affected, how likely the tool is to fail, what failure would cost, whether errors can be corrected, and whether risks can be mitigated.

Which school AI uses are considered high-risk?

Uses that affect students’ rights, records or opportunities: discipline, attendance consequences, grading, IEP goals, educator evaluation and surveillance. Georgia’s education department flags evaluating educators and writing IEP goals as high-risk, and New York City’s guidelines bar AI for counseling, grading, surveillance and student discipline, according to the CDT toolkit.

How can I tell if a school notice was made by AI?

Often you can’t from the notice itself. Ask directly which system produced it and whether a person reviewed it. If the notice affects your child’s record, ask how to request human review. Schools should be able to answer both questions clearly.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Banerji, O. (2026, October 14). “Can AI Crack the Code on Some of the Biggest Challenges in Schools?” Education Week. https://www.edweek.org/technology/can-ai-crack-the-code-on-some-of-the-biggest-challenges-in-schools/2026/10
2. Langreo, L. (2026, October 30). “Nearly 1 in 3 District and School Leaders Skip AI for Operations. Here’s What the Rest Do.” Education Week. https://www.edweek.org/technology/nearly-1-in-3-district-and-school-leaders-skip-ai-for-operations-heres-what-the-rest-do/2026/10
3. Laird, E., Woelfel, K., & Quay-de la Vallee, H. (2026, September 24). “From Possibility to Practice: A School Leader’s Toolkit for Responsible AI.” Center for Democracy & Technology. https://cdt.org/insights/from-possibility-to-practice-a-school-leaders-toolkit-for-responsible-ai/
4. Federal Trade Commission. “Complying with COPPA: Frequently Asked Questions.” https://www.ftc.gov/business-guidance/resources/complying-coppa-frequently-asked-questions
5. US Department of Education, Student Privacy Policy Office. “Protecting Student Privacy While Using Online Educational Services: Requirements and Best Practices.” https://studentprivacy.ed.gov/resources/protecting-student-privacy-while-using-online-educational-services-requirements-and-best
6. California Department of Education. (2026). “Model Policy: Artificial Intelligence in Education.” https://www.cde.ca.gov/ci/pl/aipolicy.asp