The Compensated Training Mandate: How Maryland, Idaho, and Oklahoma Are Rewriting the Rules for Teacher AI Professional Development
Three enacted state laws in 2026 move beyond “encourage training” rhetoric to require structured, paid professional development for educators learning to use AI — with federal backing and a six-week “AI dividend” as evidence the investment pays off.
The Shift From Guidance to Mandate
For two years, the dominant story in educator AI policy was a gap: teachers were told to use AI, but few were trained. As of October 2024, nearly 60% of K-12 teachers reported receiving no AI training at all. By the winter of 2025-26, that figure had fallen to 42%, with 22% reporting multiple training sessions and 9% reporting ongoing training [1]. The trend is upward, but it leaves a clear majority of educators still navigating generative AI tools with little or no formal preparation.
What changed in 2026 is not just the volume of training but its legal character. A cluster of newly enacted state laws moves the policy lever from “encourage” to “require” — and, critically, from “unpaid expectation” to “compensated obligation.” The federal government has reinforced this shift with an executive order calling for “comprehensive AI training for educators” and prioritizing discretionary grant funds for that purpose [2]. The result is a new policy model in which teacher AI professional development is becoming a funded statutory mandate rather than an optional, after-hours activity.
Three enacted laws illustrate the model in different ways:
- Idaho SB 1227 — Directs the State Department of Education to develop a comprehensive generative AI framework that explicitly includes professional development standards, AI literacy benchmarks, and data privacy requirements. Districts are required to adopt aligned policies. The law also prohibits AI from replacing human teachers [3].
- Oklahoma SB 1734 — The Responsible Technology in Schools Act requires educator-directed, human-in-the-loop AI use in all instructional contexts. It prohibits AI from serving as the primary basis for grading, discipline, or placement, mandates written district AI policies before the 2027-28 school year, and requires annual disclosure to families of every AI vendor in use [4].
- Maryland SB 720 — The Artificial Intelligence Ready Schools Act goes the furthest on the training dimension, requiring the State Department of Education to provide professional development in AI to educators and school leaders through a statewide train-the-trainer model, with the intent of training educators by July 1, 2027 [5].
The core logic unifying these laws is a rejection of the voluntary model. When AI training is optional, the teachers who show up are the ones already comfortable with technology — widening the gap between early adopters and the majority. By embedding training requirements in statute and tying district compliance to state funding, these laws attempt to close the gap systemically rather than relying on individual motivation.
The Maryland Innovation: Paying Teachers to Learn
Maryland’s Artificial Intelligence Ready Schools Act, signed by Governor Wes Moore on May 26, 2026 and effective June 1, 2026, introduces a policy feature that no other enacted state law has matched: a statutory compensation requirement for teacher AI training [5].
The law’s training provision specifies that the Department shall provide professional development in AI to educators and school leaders that is “sufficient in duration and quality to prepare the educators and school leaders” for professional use of AI, classroom application, and understanding of ethical considerations [5]. The substantive coverage required by the statute includes:
- The purpose of artificial intelligence in education
- Basic AI vocabulary and concepts
- How AI works
- Effective classroom use cases, including streamlining administrative tasks, supporting specially designed instruction, and providing timely feedback
- Effective use, privacy, security, academic integrity, and means of avoiding overdependence [5]
The training must be offered either during the employee’s regular workday as part of time designated for professional development, or as a course eligible for credit toward licensure renewal [5]. And critically, the law states that teachers in the professional development training “shall be compensated for the training with time, money, or recertification credits” and that “teachers may not be charged to attend or complete the training” [5].
The reasoning here is straightforward but consequential. Unpaid, after-hours training disproportionately burdens teachers who are already overworked — the exact population most in need of support. By requiring compensation in one of three forms (time, money, or recertification credits), Maryland ensures that the cost of AI readiness falls on the system, not on individual teachers’ personal time and wallets. This is a structural answer to the “voluntary training” problem: when the state pays, participation ceases to be a signal of personal motivation and becomes a baseline expectation.
The law also establishes a train-the-trainer model, coordinated through the Maryland Center for Computing Education and instructional coaches, with the explicit intent of training educators by July 1, 2027 [5]. This model is designed for scalability — rather than hiring external trainers for every district, the state builds internal capacity by training a cadre of educators who then train their colleagues. Each local school system must also designate an AI coordinator at the central office level, creating a designated human point of contact between the state and each district [5].
Oklahoma: Human Oversight as the Training Foundation
Oklahoma’s approach, signed by Governor Kevin Stitt on May 12, 2026, takes a different angle on the same problem. Rather than mandating a specific training curriculum, SB 1734 — the Responsible Technology in Schools Act — builds the need for educator competence into its core operational requirements [4].
The law requires that “AI tools cannot be used for instructional or educational purposes except through educator-directed use” and that “AI tools cannot be used as the primary basis for grading, discipline or promotion” [4]. These requirements are already in effect as of July 1, 2026, under an emergency clause. The policy document has until the start of the 2027-28 school year.
The implicit logic is that human-in-the-loop oversight cannot exist without trained humans. If AI tools may only be used under direct educator supervision, then educator competence becomes a precondition for legal compliance — not an optional enhancement. This creates a powerful incentive structure: districts that fail to train their teachers cannot legally deploy AI tools in instruction, which means training is not just a best practice but a compliance necessity.
Oklahoma’s law also requires annual disclosure to families of every AI vendor in use and the student data each collects, preserves parental opt-out rights for student-facing AI tools without academic penalty, and ties compliance to participation in state-funded education programs [4]. Districts that fail to meet the requirements may face loss of eligibility for state funding streams, parental enforcement actions, and scrutiny during accreditation reviews [4].
Idaho: The Framework Approach
Idaho’s SB 1227, signed by Governor Brad Little, takes the most comprehensive framework approach of the three. The law directs the State Department of Education to develop a generative AI framework addressing privacy, procurement safeguards, transparency, academic integrity, AI literacy standards, and professional development [3]. Districts are required to adopt policies aligned with the state framework.
Idaho has already begun implementation: the State Department of Education has published a professional learning calendar for 2026-2027, launched an AI Readiness Webinar Series, and opened public input on draft K-12 AI standards [3]. The framework approach gives districts flexibility in implementation while ensuring that professional development is not an afterthought but a required component of any district AI policy.
The law also prohibits AI from replacing human teachers — a guardrail that reinforces the human-in-the-loop principle embedded across all three states’ laws [3]. The logic is that if AI supplements rather than replaces instruction, then teacher competence in directing that supplementation becomes the linchpin of the entire system.
The Federal Backdrop
On April 23, 2025, President Donald Trump signed Executive Order 14277, “Advancing Artificial Intelligence Education for American Youth,” which established the White House Task Force on Artificial Intelligence Education and declared it U.S. policy to promote “comprehensive AI training for educators” [2]. The order called for prioritizing discretionary grant funds for educator AI training and directed the Secretary of Education to identify AI in education as a grantmaking priority [2].
The federal order provides rhetorical and funding support but leaves the structural design of training programs to states. This means the state laws described above are the operative policy — the federal government is supplying the funding signal, but states are defining what training looks like, who pays for it, and what happens when districts fail to comply.
The Evidence: Does Training Pay Off?
The policy case for investing in teacher AI training is supported by a growing body of survey evidence. A June 2025 Gallup-Walton Family Foundation national survey of 2,232 U.S. public K-12 teachers found that teachers who use AI tools at least weekly save an average of 5.9 hours per week — amounting to six weeks over the course of a 37.4-week school year [6]. Monthly users save an average of 2.9 hours per week [6].
The survey also found that schools with an AI policy see a 26% larger “AI dividend” — equivalent to 2.3 hours saved per week per teacher, compared with 1.7 hours in schools without such a policy [6]. This correlation between policy and benefit suggests that the state laws requiring district AI policies may amplify the time savings teachers gain from AI tools, not just the compliance burden they impose.
Teachers who use AI regularly also report quality improvements: 64% say AI improves the quality of materials adapted to meet student needs, 61% report higher-quality insights from student data, and 57% see improvements in the grading and feedback they provide [6]. Very few — 16% or less — say quality is diminished [6].
Despite these benefits, adoption remains uneven. Only 32% of teachers use AI at least weekly, 28% use it infrequently, and 40% do not use it at all [6]. Just 19% of teachers report their school has an AI policy [6]. This is the gap the new state mandates aim to close: the evidence shows AI delivers a dividend, but only for teachers who are trained, supported, and operating under clear policies.
The Compliance Challenge
These laws create real operational demands on districts. Oklahoma’s SB 1734, for example, requires districts to know what AI tools are running on student devices, maintain human oversight in instructional contexts, and disclose that information to families annually [4]. A generic technology use policy will not satisfy the law — the policy must specifically address AI tools, their instructional role, and the boundaries around high-stakes uses like grading and discipline [4].
For districts, the immediate action items include:
- Audit current AI tools — know what is active on student devices before drafting any policy [4].
- Review vendor contracts for data collection and retention terms, since annual disclosure to families is now required [4].
- Set up parental opt-out processes — the right to opt out of student-facing AI tools without academic penalty is already active in Oklahoma [4].
- Build training timelines — Maryland’s train-the-trainer model targets July 1, 2027; Oklahoma’s written board-adopted policy is due before the 2027-28 school year [4][5].
The deeper challenge is that training quality matters as much as training quantity. As Education Week reported in May 2026, the increase in AI professional development has been real but uneven — “from the frequency of training to the type and depth of information imparted” — and many teachers remain reluctant to learn about using AI in their work [1]. Mandating training does not guarantee engagement; the structure and compensation model determines whether teachers actually absorb what is taught.
What Comes Next
The 2026 legislative session produced 77 tracked AI-in-education bills across 27 states, with 10 enacted so far [7]. Maryland, Oklahoma, and Idaho represent the leading edge, but other states are pursuing parallel paths. South Carolina’s H.B. 5253 would require written parental opt-in consent, annual public disclosure of AI tools, and prohibit AI from replacing licensed teachers in core instruction or grading [7]. West Virginia’s H.B. 5205 would require the State Board to develop model AI policies that automatically apply to districts that fail to adopt compliant local policies by July 2027 [7].
The convergence point across these efforts is clear: AI in classrooms is no longer a question of whether, but of how — and the “how” is being defined by laws that require trained, compensated educators to remain at the center of the instructional relationship. The states that have moved first are betting that paying teachers to learn AI is cheaper and more effective than deploying AI without them.
References
[1] Education Week — “More Schools Are Providing AI Training for Teachers. Is It Any Good?” by Elizabeth Heubeck (May 18, 2026) [SECONDARY]
[2] The White House — “Advancing Artificial Intelligence Education for American Youth,” Executive Order 14277 (April 23, 2025) [PRIMARY]
[3] Idaho State Department of Education — “Artificial Intelligence” (2026) [PRIMARY]
[4] Lightspeed Systems — “Oklahoma SB 1734 Is Now Law: What Every District Needs to Do Before 2027-28” by Larissa Somaio (July 22, 2026) [SECONDARY]
[5] Maryland General Assembly — Chapter 634 (Senate Bill 720), Artificial Intelligence Ready Schools Act, approved May 26, 2026 [PRIMARY]
[6] Walton Family Foundation — “The AI Dividend: New Survey Shows AI Is Helping Teachers Reclaim Valuable Time” (June 25, 2025) [ADVOCACY]
[7] FutureEd — “Legislative Tracker: 2026 State AI in Education Bills” by Bella DiMarco (Updated July 13, 2026) [SECONDARY]
AI automated and generated content. Human reviewed.
The Weekly AI Policy Brief
AI policy moves fast. Get the week’s developments, new trackers, and podcast episodes in one short email — free, ad-free, no spam.
Leave a Reply