Washington, D.C. public and charter school teachers just got their first detailed rulebook for using artificial intelligence in the classroom, and it draws a hard line around a few high-stakes decisions: no AI grading students, no AI handling discipline, and no AI weighing in on individualized education plans for students with disabilities. The District's Office of the State Superintendent of Education issued the new AI policy for the 2026-2027 school year, giving teachers a framework for where the technology is welcome and where it is not.
The guidance arrives as a non-mandatory model policy rather than a binding mandate, according to WTOP News. At its center is a stoplight-style system developed by Antoinette Mitchell to help educators quickly judge when and how AI belongs in their work, sorting tasks into categories similar to red, yellow and green signals. Teachers can lean on AI for drafting lesson plans, scheduling, analyzing attendance data, analyzing assessment data, and drafting messages to parents, per the same account. But AI-assisted grading and review of student work fall into a use-with-caution category, meaning substantial human oversight is required, and Mitchell said those calls should ultimately be made by people, not algorithms.
Where the Red Lines Are Drawn
The policy is unambiguous about several areas where AI has no role at all. AI use for determining individualized education plans for students with disabilities is prohibited outright, as is AI involvement in student discipline and in teacher performance evaluations. Physical surveillance by AI is also barred, along with any AI use that could compromise student safety or privacy, the outlet reports. AI monitoring of digital activity on school-issued devices is permitted, but only with substantial human oversight built in.
Mitchell explained her thinking on grading specifically, telling the outlet that feedback was critical to her own growth and development in school. She said grading should generally rest on a teacher's own judgment rather than an algorithm's output, and that in the yellow, use-with-caution zones, humans must make the final call whenever AI is involved. Teachers are also restricted to platforms and tools their district has actually approved, rather than picking their own AI software.
Why D.C. Schools Asked for This
The push for clearer rules did not come out of nowhere. D.C. schools and administrators had specifically requested this kind of guidance, and survey data backs up why: a February 2026 survey by the Office of the State Superintendent of Education and Mathematica's Regional Educational Laboratory Mid-Atlantic found that only 45% of responding D.C. school districts had staff AI policies and just 34% had student AI policies. District leaders in that same survey named model instructional guidance as their single highest-priority need.
The rollout coincides with new online professional development courses for local educators, titled AI Literacy for Educators and Instructional Planning with AI, alongside the model policy's release on Tuesday. Districts are now expected to provide robust AI training for teachers as they put the new framework into practice. The guidance also connects to a broader legislative push: the D.C. Council introduced Bill 26-0491, the Artificial Intelligence Literacy in Education Act of 2025, in November 2025, directing local education agencies serving high school students to establish formal AI literacy policies by the start of the 2026-2027 academic year, according to BillTrack50.
A Regional and National Pattern
D.C. is not acting in isolation. Neighboring Maryland and Virginia both enacted 2026 legislation, Maryland's S.B. 720 and Virginia's S.B. 394, instructing their state education departments to issue statewide K-12 AI frameworks and educator training guidelines, as detailed by MultiState. Nationally, the pattern holds: as of December 2025, 35 states and territories had published official K-12 AI guidance, but the National Association of State Boards of Education notes that virtually all of it remains nonbinding, leaving final implementation to local school boards, much like D.C.'s own approach.
The stakes behind the paperwork are real. An August 2026 national survey of 1,017 teenagers by Common Sense Media's Youth AI Safety Institute found that 70% of U.S. teens use AI for schoolwork, yet only 27% have ever discussed what AI is and how it works with a classroom teacher. Separately, TechPolicy.Press reported in August 2026 that 60% of K-12 educators surveyed nationwide found their district's AI policies unclear, a confusion gap OSSE's new stoplight system is explicitly meant to close.
Even as adoption accelerates, some researchers are flagging a downside. A March 2026 RAND Corporation study surveying 1,214 young people ages 12 to 29 found that while AI homework assistance is rising, a growing share of youth worry that relying on AI tools is eroding their critical thinking skills. That tension, between convenience and cognitive cost, is part of why OSSE built the new policy into its five-year strategic plan, Building Momentum for Every Student's Future: 2026-2030, published in June 2026, which frames AI readiness as central to preparing students for future work.
What Comes Next for Teachers
Because the policy is not mandatory, how consistently it gets applied may vary from one D.C. campus to the next, with individual districts and even individual teachers still filling in gaps as they go. The professional development courses launched alongside the policy are meant to help close that gap, but for now, frontline teachers remain the ones deciding, case by case, where AI earns their trust and where a human judgment call still has to win out.
