From Guidance to Mandates: The New Map of K-12 AI Policy in U.S. States, October 2026

The Patchwork Is Becoming Law

Two years ago, AI policy in American schools mostly meant memos and guidance documents. An updated state-by-state tracker published this month by Progress Learning shows how fast that is changing: K-12 AI governance is hardening into statutes, deadlines and — in a growing number of states — enforceable mandates that district leaders can no longer treat as optional reading.

Five Regulatory Modes

The tracker, updated October 2026, sorts state approaches along a spectrum from lightest to strictest:

  • Permissive guidance — Georgia, Illinois (ISBE guidance released July 2026), Massachusetts, North Carolina (model policy due December 2026), Oregon and Washington offer frameworks and rubrics, leaving adoption to districts.
  • Managed integration — Ohio's House Bill 96 required every district to adopt a formal AI policy by July 1, 2026; Tennessee's Public Chapter 550 mandates board-adopted rules on student and staff AI use; Utah's HB 273 has a July 2027 deadline; Virginia's SB 394 pairs statewide guidance with required professional development.
  • Mandated oversight — California's AB 1159 bans using identifiable student data to train AI models and requires notification and consent for AI grading; Maryland's SB 720 (the AI Ready Schools Act, effective June 2026) gives districts 120 days to adopt policies and bans high-stakes automated decisions; New York's proposed A9190 would prohibit classroom AI before ninth grade.
  • Direct parent safeguards — Florida's Rule 6A-1.0957 (adopted September 2026) requires parent opt-in for student-facing AI tools, public approved registries and interaction logs; Oklahoma's SB 1734 lets parents opt out of student-facing tools without academic penalty.
  • IT and data-privacy baseline — Alabama anchors procurement in the NIST AI Risk Management Framework; Idaho's SB 1227 targets AI edtech tools explicitly. And 33 states still have no specific instructional AI mandate, relying on FERPA and COPPA baselines.

The Federal Signal

Hovering over the patchwork is the U.S. Department of Education's August 20, 2026 guidance, which pushes districts toward a more nuanced evaluation of classroom technology: judge a tool by the specific learning problem it solves, for whom, and for how long — treating AI as a precision instrument rather than passive “computer time.”

What District Leaders Should Take Away

Three implications stand out. First, the compliance surface is expanding fast — policies adopted this school year must account for training-data bans, opt-in regimes and logging duties that differ dramatically across state lines. Second, instructional efficacy is becoming the legal standard: policies increasingly ask whether a tool demonstrably supports learning, not merely whether it is safe. Third, the 33-state baseline group is the one to watch — if 2027 legislative sessions follow the Ohio-Maryland-Florida pattern, the map will look very different a year from now.

Finns också på: English Deutsch Español