AI Regulation Tracker / State law
Oklahoma law requires every school district to adopt a written AI policy before the 2027-28 school year
Oklahoma SB 1734, the Responsible Technology in Schools Act, directs every public school district to adopt a written artificial intelligence policy and to keep AI use educator-directed with a human in the loop. Districts, ed-tech vendors, and K-12 privacy officers now have a fixed deadline and a defined minimum standard.
Oklahoma has set a firm deadline for a task many school districts have handled informally until now. Under Senate Bill 1734, the Responsible Technology in Schools Act, every public school district board of education must adopt and maintain a written policy governing the use of artificial intelligence and emerging technologies before the beginning of the 2027-2028 school year. The bill passed the Senate 42-0 and the House 89-0 and was signed into law, according to legislative tracking, on May 12, 2026. It is codified as a new section of the Oklahoma Statutes at Section 11-120 of Title 70.
The law is short, but it does two things at once. It sets statewide guardrails that apply to any district that uses AI, and it requires each district to write those guardrails into a local policy with a defined minimum content. For administrators, vendors, and privacy officers, the practical question is no longer whether to have an AI policy. It is whether the policy on file meets the specific elements the statute names.
What the statute requires of AI use
The act applies whenever a public school district uses artificial intelligence, and it attaches several conditions. Classroom AI tools must be deployed through what the law calls "educator-directed AI use," and the statute is explicit that responsibility for decisions informed by AI "shall remain with school employees and shall not be delegated to an artificial intelligence system or vendor."
Both classroom and student-facing AI tools must operate with a human in the loop. The law defines "human-in-the-loop" to mean that an educator or authorized school employee "maintains oversight, review, and final decision-making authority over outputs generated by an artificial intelligence tool." Outputs generated by AI must be reviewed by an educator or authorized employee before they are used in instruction, feedback, assessment, or decision-making.
The act also draws a hard line on high-stakes decisions. AI tools "shall not be used as the primary basis for student grading, discipline, placement, promotion, retention, or other high-stakes educational decisions." Student-facing tools must be appropriate to the age and developmental level of students and used only for clearly defined educational purposes. On data, the law requires compliance with applicable federal and state privacy and security laws, including the Family Educational Rights and Privacy Act of 1974 (FERPA), and directs districts to "take reasonable steps to minimize the amount of student data shared with artificial intelligence tools."
The written policy each board must adopt
The centerpiece for administrators is the policy mandate. The statute states that "prior to the beginning of the 2027-2028 school year, each school district board of education shall adopt and maintain a policy governing the use of artificial intelligence and emerging technologies." At a minimum, that policy must do five things and comply with a sixth.
It must identify the roles or school employees responsible for approving and overseeing the use of AI tools. It must address appropriate instructional uses and prohibited uses. It must address student data protection and data minimization practices. It must address transparency to students and families. And it must provide for periodic review and updates. The sixth requirement ties the local policy back to the statewide floor: the policy must comply with the substantive use conditions in the act and with the guidance the State Department of Education develops.
That last point matters for timing. The State Department of Education is directed to "develop and make available guidance regarding the responsible use of artificial intelligence and emerging technologies for public school districts." A district policy has to be consistent with that guidance, so districts that draft in isolation risk having to revise once the state guidance appears. The State Board of Education may also promulgate rules to implement the section.
Transparency and the parental opt-out
The law gives families two concrete rights. First, each district must provide parents or legal guardians a written disclosure "not less than annually." That disclosure has to identify all AI tools in use in the district, the vendors providing them, the categories of student data collected, the extent to which student data is shared, and the educational purpose for which each tool is used. This is the provision ed-tech vendors will feel most directly, because a district cannot complete the disclosure without vendor cooperation on data categories, data sharing, and purpose.
Second, a parent or legal guardian "may provide written notice opting the student out of participating in student-facing AI tools at any time." Students who opt out "shall not be academically penalized or denied access to core instructional content." Districts will need a workable process for honoring opt-outs without cutting a student off from required coursework.
What the law does not do
The act is deliberate about its limits. It states that nothing in the section shall be construed to require a public school district to use artificial intelligence in any school, or to limit the authority of district boards to make decisions about instructional methods or technology use. The legislative findings frame the same idea in plain terms: classroom teachers play an "essential and irreplaceable role," and AI tools "should supplement educator-led instruction, not supplant it." The law is a floor and a process requirement, not a mandate to adopt AI and not a ceiling on local discretion above the statutory minimum.
How to build the policy that the statute demands
For a district compliance officer, the fastest way to meet the mandate is to work backward from the six required elements and the use conditions in the act. Start with an inventory. The annual disclosure cannot be written, and the policy cannot define approved and prohibited uses, until the district knows which AI tools are already in classrooms, who introduced them, and what student data each one touches. Many tools now arrive inside general-purpose software, so the inventory should look past products labeled as AI and capture features that generate content, feedback, or assessment.
Next, assign ownership. The statute requires the policy to identify the roles or employees responsible for approving and overseeing AI tools, which means a named person or office, not a general reference to staff. That owner should control a short approval path: a tool is not classroom-ready until someone has confirmed it meets the age-appropriateness, data-minimization, and human-review conditions the law sets. Because the policy must provide for periodic review, the same owner should hold a recurring calendar item to revisit the tool list and the disclosure at least once a year.
Vendors have a parallel task. A district cannot complete its annual parent disclosure without the categories of student data a tool collects, the extent to which that data is shared, and the educational purpose it serves. Vendors that package those three answers into a standard data sheet, and that can show how a human stays in control of any output used for grading or discipline, will move through district approval faster than those that cannot. The FERPA and data-minimization language in the act gives procurement teams a concrete basis to ask for that documentation before a contract is signed.
One more sequencing point deserves attention. Because each local policy must comply with the State Department of Education guidance, districts that draft everything before the guidance is published may have to amend later. A workable approach is to build the structural pieces now, the inventory, the named owner, the approval path, and the disclosure template, then finalize the substantive use rules once the state guidance is available.
Why this reaches beyond Oklahoma
For ed-tech vendors and multistate compliance teams, Oklahoma is one more state converting general expectations about AI in schools into a datable, auditable obligation. The disclosure elements, the human-in-the-loop requirement, and the ban on AI as the primary basis for high-stakes decisions echo themes appearing in other 2026 state education bills. A vendor that can already document data categories, data sharing, educational purpose, and human oversight for one state is better positioned for the next. Districts operating near state lines, and companies selling across them, should treat the Oklahoma elements as a practical template rather than a local exception.
Frequently Asked Questions
What does Oklahoma SB 1734 require school districts to do?
It creates the Oklahoma Responsible Technology in Schools Act and requires each public school district board of education to adopt and maintain a written AI policy before the 2027-2028 school year. The policy must, at minimum, name the staff responsible for approving and overseeing AI tools, address appropriate and prohibited instructional uses, address student data protection and minimization, address transparency to students and families, and provide for periodic review, all consistent with State Department of Education guidance.
Who is affected by the law?
Oklahoma public school districts and their boards, classroom teachers and staff who direct AI use, K-12 compliance and student-data-privacy officers who will write and audit the policy, and ed-tech vendors, who must supply the data-category, data-sharing, and educational-purpose details that districts need for the required annual parent disclosure.
When does the law take effect and when must policies be in place?
The act carries an emergency clause and a stated effective date of July 1, 2026, and was signed, according to legislative tracking, on May 12, 2026. Each district board must adopt and maintain the required written policy before the beginning of the 2027-2028 school year.
Can AI be used to grade students or make discipline decisions?
Not as the primary basis. The statute states that AI tools shall not be used as the primary basis for student grading, discipline, placement, promotion, retention, or other high-stakes educational decisions, and it requires a human in the loop, meaning an educator or authorized employee keeps oversight, review, and final decision-making authority over AI outputs.
Can parents opt their child out of AI tools?
Yes. A parent or legal guardian may provide written notice at any time opting the student out of student-facing AI tools, and the district may not academically penalize the student or deny access to core instructional content. Districts must also provide a written disclosure to families at least annually listing the AI tools in use, the vendors, the student data collected and shared, and the educational purpose of each tool.
Sponsored Training
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.