Current status: Last reviewed September 7, 2026. This resource summarizes the enacted act as published by the Virginia Legislative Information System and cited public sources. It is not legal advice. Confirm obligations with division counsel and the Virginia Department of Education.
Virginia now has a school AI statute, and it works through guidance rather than a checklist. House Bill 1186 and Senate Bill 394, identical companion bills, were approved on April 13, 2026 as Chapters 937 and 943 of the 2026 Acts of Assembly and took effect July 1, 2026. Together they created Code of Virginia section 22.1-20.2:1.
The section gives the Department of Education two jobs and every local school board one. The Department must compile information on current uses of AI systems for student instruction and publish guidance on a public web page. Each school board must then establish, implement, and enforce policies consistent with that guidance. The statute sets no calendar deadline for either step, and the enacted text does not contain the “no required chatbot” language that circulated during the session.
The law at a glance
| Question | Answer from Chapter 943 |
|---|---|
| What is the source? | 2026 Acts of Assembly, Chapters 937 (HB 1186) and 943 (SB 394), creating Code of Virginia section 22.1-20.2:1 |
| When did it take effect? | July 1, 2026 |
| What must the Department do? | Compile information on current uses of AI systems for student instruction, and establish and post guidance in a publicly accessible location on its website |
| What must school boards do? | Establish, implement, and enforce policies consistent with the Department’s guidance |
| Deadlines | None in the enacted text for the guidance or for board policies |
| Pilot | An AI Innovation in Education Pilot Program, run with appropriated funds, priority to high-poverty, rural, and under-resourced divisions; the pilot subsection expires July 1, 2030 |
| Funding | A budget amendment provided $2,000,000 in general funds for fiscal year 2027 for the pilot |
| Votes | House 95-0 on both bills; Senate 38-1 and 37-2, as reported by VPM |
The act defines an “artificial intelligence system” broadly as a machine-learning-based system that infers from inputs for an explicit or implicit objective, excluding systems still in pre-release research and development. It also defines an AI system used in instructional settings, a division-managed AI platform, and a division-managed environment. Those last two terms matter because the guidance must favor them.
What the guidance must address
Subsection B enumerates what the Department’s guidance must cover. A school board that wants to be ready before the guidance lands can organize its policy work around the same list.
| Statutory element | What a division policy will need to say |
|---|---|
| Student data privacy, including FERPA and COPPA | Which student data may reach which AI system, under what authority, and how records are handled |
| Resources and training for teachers | What professional development is offered, to whom, and how completion is tracked |
| Transparency and explicability | How students, families, and staff learn that AI is in use and what it does |
| Data-processing agreements that prohibit use of division-level or student data for training or improving external AI models | Contract language for every AI vendor, including AI features inside existing products |
| Best practices on bias and discrimination, protocols and professional development, prohibitions on relying solely on an AI system for certain high-stakes decisions, accessibility, and equity | Human-review rules for grading, placement, discipline, and similar decisions; accessibility review in tool approval |
| Alignment with the Standards of Learning | Instructional use cases mapped to standards rather than adopted as general-purpose tools |
| Guardrails that prioritize division-managed AI platforms | A preference, in policy and procurement, for AI reached through division-managed environments over consumer accounts |
| Teacher access to transcripts of student interactions with an AI system, plus dashboards, automated alerts, and audit logs | The technical and contractual ability for a teacher to see what a student and an AI system exchanged, and for the division to audit it |
The transcript-access element is the one most divisions cannot satisfy today with consumer AI accounts. A teacher can see a student’s Google Doc; a teacher generally cannot see a student’s conversation in a consumer chatbot. The statute’s preference for division-managed platforms follows from that gap.
What the enacted text does not say
Avoid restating session coverage as law. The act as published:
- sets no date by which the Department must publish guidance;
- sets no date by which school boards must adopt policies;
- does not prohibit a school from requiring a student to use a chatbot (news reports describe that language as dropped in the compromise version, and it does not appear in Chapter 943);
- does not name a required product, vendor, or platform;
- does not itself create a parent opt-out; and
- does not apply to private schools.
Other Virginia and federal law can still apply to an AI use, including student-records, procurement, civil-rights, accessibility, and cybersecurity duties, and the 2024 Executive Order 30 guidelines remain the Commonwealth’s current published AI-in-education guidance until the Department issues the guidance this section requires.
The pilot program
Subsection D authorizes the AI Innovation in Education Pilot Program with such funds as are appropriated. The Department must give priority to high-poverty, rural, and under-resourced school divisions. Participating divisions must provide educator professional development and evaluation plans. The Department must report to the chairs of the House Committee on Education and the Senate Committee on Education and Health by December 1 of each year. The subsection expires July 1, 2030.
A budget amendment provided $2,000,000 in general funds for fiscal year 2027 and none for fiscal year 2028. Divisions interested in the pilot should watch for the Department’s application process alongside the guidance.
What a school board can do now
The guidance has no due date, so a board that waits for it may wait a while. The statutory list is public, and the records it implies are the same records every state’s AI law eventually asks for.
- Inventory AI systems in instructional settings, including AI features embedded inside approved products. The Department is compiling the same inventory statewide.
- Reconcile existing policies: acceptable use, academic integrity, student records, procurement, and professional development. Note where each statutory element already has a home.
- Add the training-prohibition clause to every AI data-processing agreement now. It is the one contractual requirement the statute states directly.
- Decide the division-managed platform strategy. Which AI capabilities will students reach through division-managed environments, and which consumer accounts will the division stop relying on?
- Test transcript access and audit logs for each approved student-facing tool. If a teacher cannot see the interaction and the division cannot audit it, the tool will be hard to keep once the guidance is published.
- Set the human-review rule for high-stakes decisions in writing, and name the decisions it covers.
- Prepare the board record: owners, procedures, evidence plan, and review cadence, so adoption can follow the guidance quickly.
How Tenet may support a Virginia implementation
Tenet by TrueMadeAI is K-12 AI governance software. Tenet Edge applies district and classroom policy on supported direct-use AI surfaces on managed Chrome. Tenet Gateway is a founding-district program for approved backend AI operations.
A division implementation could use Tenet to help:
- present role-, class-, subject-, and schedule-aware rules on supported surfaces, which supports the transparency element;
- apply supported on-device data-loss-prevention controls before student text leaves the device, which supports the student-data-privacy element;
- route approved backend AI operations through a division-managed path with defined application, purpose, data, model, and audit boundaries, which supports the preference for division-managed platforms and the audit-log element; and
- retain bounded operational evidence for review without treating raw student conversations as analytics.
Tenet does not replace the Department’s guidance, board policy adoption, procurement, educator judgment, or the division’s responsibility to determine which uses the section covers. Transcript access for a teacher depends on the specific tool and contract. Review the dated capability matrix before relying on a specific product surface.
If your division wants to map the statutory guidance elements to owners, records, and supported controls before the guidance is published, request a scoped Tenet District conversation.
What this guide does not establish
This guide does not determine:
- when the Department will publish its guidance or what it will say;
- whether a particular division use is an AI system in an instructional setting under the act;
- whether other state or federal law creates additional consent, privacy, accessibility, or records duties;
- that a contract, an approved list, or a technical product alone creates compliance; or
- the legal consequence of adopting a policy before or after the guidance appears.
Those conclusions require the current official sources, the division’s facts, and qualified review. The state laws and guidance tracker places Virginia alongside the other states with binding duties.
Sources
- Chapter 943, 2026 Acts of Assembly (SB 394), enrolled act text
- Code of Virginia section 22.1-20.2:1
- HB 1186 (2026) bill details
- SB 394 (2026) bill details
- Budget amendment Item 124 #5s, AI Innovation in Education Pilot Program
- Guidelines for AI Integration Throughout Education in the Commonwealth of Virginia, Executive Order 30
- VPM, AI guidance for schools bills head to the governor, March 25, 2026
Frequently asked questions
When does Virginia’s school AI law take effect?
July 1, 2026. HB 1186 and SB 394 were approved on April 13, 2026 as Chapters 937 and 943 of the 2026 Acts of Assembly and created Code of Virginia section 22.1-20.2:1.
Does Virginia set a deadline for school boards to adopt an AI policy?
No. The enacted text requires each school board to establish, implement, and enforce policies consistent with the Department of Education’s guidance, but it sets no calendar date for the guidance or for board adoption. Boards should track VDOE’s publication date and plan adoption from there.
Does the Virginia law prohibit schools from requiring students to use an AI chatbot?
The enacted text does not contain a chatbot provision. Reporting during the session described such language in an earlier version. Districts should rely on the act as published and on the VDOE guidance once it is released.
What must the Virginia Department of Education’s AI guidance address?
Among other things: student data privacy including FERPA and COPPA; resources and training for teachers; transparency and explicability; data-processing agreements that prohibit using division or student data to train external AI models; best practices on bias, discrimination, professional development, high-stakes decisions, accessibility, and equity; and alignment with the Standards of Learning. It must also address guardrails that prioritize division-managed AI platforms and teacher access to transcripts, dashboards, alerts, and audit logs.
What is the Virginia AI Innovation in Education Pilot Program?
A pilot the Department may run with appropriated funds, with priority for high-poverty, rural, and under-resourced divisions. Participants must provide educator professional development and evaluation plans, and the Department must report annually by December 1. The pilot subsection expires July 1, 2030. A budget amendment provided $2,000,000 in fiscal year 2027.
Does Virginia require a specific AI product or vendor?
No. The law directs guidance and local policy. The guidance must prioritize division-managed AI platforms and require data-processing agreements that bar training on division or student data, which affects how any product is contracted, not which product is chosen.