Table of Contents
School AI Oversight: Virginia Is Auditing the Rules
Virginia's legislative audit commission is studying college AI policy. The four-part school AI oversight framework it uses also works on any school board.
School AI oversight has mostly meant districts writing their own rules and grading their own homework. On September 24, 2026, Virginia’s Joint Legislative Audit and Review Commission began assembling a team to study AI policies across the state’s public colleges and universities, covering academic integrity, data privacy, equity and access, and transparency. The study binds nobody and will not report this school year. Its value to a parent is immediate anyway: a legitimate oversight body has just published the four questions worth asking about any institution’s AI rules, and those four questions work at a middle school board meeting.
Key Takeaways
- JLARC is an oversight body, not a regulator. Its own site says it “conducts program evaluation, policy analysis, and oversight of state agencies” for the Virginia General Assembly.
- The study covers Virginia’s public colleges and universities only. Not K–12, not private institutions, not other states.
- The four review areas are the deliverable worth copying: academic integrity, data privacy, equity and access, transparency.
- Expect findings in 2027, not this term. JLARC’s published studies are dated by delivery month, and recent ones span roughly a year.
- The study was not yet listed on JLARC’s own study page when checked, which is normal for a newly assembled team and worth knowing before you go looking.
What JLARC is, and what it can and cannot do
A legislative audit commission is a body that evaluates how state agencies and programmes are working and reports its findings to the legislature. It investigates. It does not regulate.
JLARC’s own description is that it “conducts program evaluation, policy analysis, and oversight of state agencies” on behalf of the Virginia General Assembly. Its recent study list gives a sense of its range: a K-12 School Accountability System review dated December 2025, Community Colleges and Higher Education Trends dated October 2025, Correctional Education Programs dated November 2025, and an Agricultural Cost-Share Program review dated September 2026.
Two things follow. First, JLARC studies education regularly and seriously, including K–12. Second, its output is a report with findings and recommendations that the General Assembly may act on, amend or ignore. If you are waiting for this study to change a rule, you are waiting for a second step that may never come.
The AI-in-education tracker entry dated September 24, 2026 describes JLARC assembling a team to study higher-education AI rules, with the review covering academic integrity, data privacy, equity and access, and transparency. One honest note: JLARC’s own study page did not list the AI review when checked, which is expected for a study in the team-assembly stage. Treat the September 24 entry as the dated record and watch the commission’s list for the study to appear with a target month.
On timing, a reasonable inference rather than a published fact: because JLARC’s studies are dated by the month they are delivered and recent ones span roughly a year from authorisation to report, a study beginning in September 2026 would plausibly report in late 2027. Plan accordingly and do not expect answers this spring.
Why a college study matters for a middle schooler’s parent
Three reasons, and the third is the one that changes what you do tomorrow.
Public universities are state agencies, which makes them auditable in a way that districts often are not. A state legislature can direct a comprehensive review of every public institution’s AI policy in one motion. Getting the same scrutiny across 130 school districts requires 130 conversations.
Findings travel. When a state oversight body publishes that, say, institutions have inconsistent disclosure practices, that language appears in the next legislative session’s K–12 bills. Higher education is frequently the testing ground for accountability language that later reaches schools.
And the framework is portable. This is the practical gift. Four areas, each with a specific question a parent can ask, and each with a recognisable bad answer.
| Review area | The question to ask your school | What a weak answer sounds like |
|---|---|---|
| Academic integrity | ”What is the written rule for AI use per assignment, and who decides?" | "We trust our teachers’ judgement” |
| Data privacy | ”Which vendors process student data, and what is the retention schedule?" | "Everything is FERPA compliant” |
| Equity and access | ”Which students get the licensed tool and which get the free tier?" | "All students have access to technology” |
| Transparency | ”Where is this published, and when was it last updated?" | "It’s in the handbook somewhere” |
Those four right-hand answers are not bad faith. They are what an under-resourced institution says when nobody has done the work. Asking the left-hand question in public, at a recorded meeting, is how the work starts.
What each of the four areas actually covers
Worth unpacking, because the labels are vaguer than the substance.
Academic integrity is not mainly about catching cheating. It is about whether the rules are stated at the level where the decision happens. A school-wide policy saying “AI use must be disclosed” does nothing for a student deciding whether using a grammar checker on a history essay counts. The functional unit is the assignment, and the useful artefact is a line on the assignment sheet.
Data privacy is where federal law sets a floor. The U.S. Department of Education describes FERPA as giving parents access, amendment and disclosure-control rights over education records under 34 CFR Part 99, with those rights transferring to the student at 18 or on postsecondary enrolment, which is precisely why a college study is about students’ own rights rather than parents’. The FTC’s COPPA Rule, whose current text was published April 22, 2025, covers under-13s and so reaches elementary and middle school rather than college. Different ages, different statutes, same vendor.
Equity and access is the one most often answered dishonestly, and the dishonesty is usually unintentional. “All students have access” typically means all students can reach something. It rarely means all students reach the configured, supervised version. That distinction has evidence behind it: Bastani and colleagues found in PNAS in 2025 that unguarded GPT-4 access reduced high school students’ retained math skill, while the same model configured to give teacher-designed hints did not. A free consumer chatbot is the unguarded condition. Equity of access to the worse version is not equity.
Transparency is the cheapest to fix and the most neglected. A policy that exists but cannot be found is functionally absent. Ask for the URL. If the answer is a PDF attached to a board packet from eighteen months ago, you have your finding.
How to run this audit on your own school
Ask the four questions in writing, one email
Not four emails. One, with four numbered questions, addressed to the principal and copied to the district technology director. Numbered questions get numbered answers; paragraphs get a phone call that leaves no record.
Ask for the vendor list before the policy
The policy describes intent. The vendor list describes what is running. Most districts maintain an approved software inventory for state reporting, and it is usually obtainable. Our walkthrough of reading a district AI policy in 20 minutes covers what to look for once you have both documents.
Check the equity question at the tool level
For each tool your child uses, ask whether it is the licensed version or the free tier. If some classes have the licensed version and others do not, you have found an in-building equity gap, which is both more fixable and more common than the between-district kind. Our piece on the AI access gap along income lines covers the larger pattern.
Use UNESCO’s framework for the curriculum half
Oversight covers rules; it does not cover what is taught. UNESCO’s 2024 AI Competency Framework for Students defines 12 competencies across four dimensions at three levels, and asking which dimensions a school’s AI instruction covers is the curriculum equivalent of JLARC’s audit question.
What not to do
Do not treat the launch of a study as a result. A commission assembling a team in September 2026 has findings in 2027 at the earliest, and in the meantime nothing about your child’s school has changed. The mistake is to read oversight news as reassurance. The correct reading is that someone has published a good set of questions, and nobody is going to ask them about your school except you.
What to Watch For Over the Next 3 Months
- Week 4: Watch for the study to appear on JLARC’s published list with a target delivery month. That is the signal it has moved from team assembly to active work.
- Month 2 red flags: A school citing “state review” as a reason to defer its own policy work. Any claim that a district is “compliant” with a study that has not reported. An equity answer that counts access without naming the tier.
- Month 3 self-check: Do you have written answers to all four questions for your child’s school? Four answers on one page. That page is more oversight than most schools currently receive.
Frequently Asked Questions
What is JLARC and what power does it have?
Virginia’s Joint Legislative Audit and Review Commission, which conducts programme evaluation, policy analysis and oversight of state agencies for the General Assembly. It investigates and recommends; it does not issue regulations. Any binding change would require legislative action afterwards.
Does this study cover K–12 schools?
No. The review covers Virginia’s public colleges and universities. JLARC does study K–12 separately, including a School Accountability System review dated December 2025, but this AI study is higher education.
When will there be findings?
Not this school year. JLARC’s published studies are dated by delivery month and recent ones span roughly a year from authorisation, so a study beginning in September 2026 would plausibly report in late 2027. That is an inference from the pattern rather than an announced date.
What are the four areas being reviewed?
Academic integrity, data privacy, equity and access, and transparency. That list is the most portable part of the story, because the same four questions work on any school at any level.
Why isn’t the study on JLARC’s website?
It was not listed when checked, which is normal for a study still in the team-assembly stage. Newly authorised studies typically appear on the commission’s list once they have a target delivery month.
My state has nothing like this. What then?
Run the four-question audit yourself on your child’s school. A single email with four numbered questions, plus a request for the vendor list, produces more written accountability than most districts generate on their own.
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
- Pursuit. (2026). “AI in Education: News, Policies, Innovations” (entry dated September 24, 2026). https://www.pursuit.us/news/ai-in-education-news-policies-innovations
- Joint Legislative Audit and Review Commission, Commonwealth of Virginia. “About JLARC and study list.” https://jlarc.virginia.gov/
- U.S. Department of Education, Student Privacy Policy Office. “What is FERPA?” 34 CFR Part 99. https://studentprivacy.ed.gov/faq/what-ferpa
- Federal Trade Commission. (2025, April 22). “Children’s Online Privacy Protection Rule (COPPA).” FTC Legal Library. https://www.ftc.gov/legal-library/browse/rules/childrens-online-privacy-protection-rule-coppa
- Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). “Generative AI without guardrails can harm learning: Evidence from high school mathematics.” Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.2422633122
- UNESCO. (2024). “AI Competency Framework for Students.” https://www.unesco.org/en/articles/ai-competency-framework-students
- Sidoti, O., Park, E., & Gottfried, J. (2025, January 15). “About a quarter of U.S. teens have used ChatGPT for schoolwork — double the share in 2023.” Pew Research Center. https://www.pewresearch.org/short-reads/2025/01/15/about-a-quarter-of-us-teens-have-used-chatgpt-for-schoolwork-double-the-share-in-2023/
