74% of Schools Have AI Policies. Most Teachers Still Have No Training.
Schools are racing to put AI policies on the books. But the people who actually have to implement them — the teachers — are being left to figure it out alone.
📊 The Gap in Numbers
A Carnegie Learning national survey found that 74% of school districts now have some form of AI policy in place. That’s a dramatic increase from just a year ago, when most districts were still debating whether to acknowledge AI at all.
But here’s the problem: the EdWeek Research Center found that 58% of teachers have received no AI training whatsoever — two full years after ChatGPT’s release. Only 43% of teachers said they’d received at least one training session on AI, and that’s a nearly 50% increase from the previous survey where just 29% said the same.
Progress? Yes. Enough? Not even close.
As Tara Natrass, managing director of innovation strategy for ISTE+ASCD, put it: “If 58 percent of teachers still have no training two years after the release of ChatGPT, then districts have a lot of work to do to get everyone up to speed.”
🪤 The Policy Vacuum
The 74 reported on a survey of state board of education members that found most states are still in the early stages of addressing generative AI. While many states provide guidance or toolkits, local decisions dominate the landscape — each school district is primarily responsible for shaping its own plan.
“We are a ‘local control’ state, so some school districts have banned [generative AI],” wrote one respondent. “Our department of education has an AI tool kit, but policies are all local,” wrote another.
The result: a patchwork of policies that varies wildly from district to district, with no baseline standards and no consistency in what teachers are expected to know or do.
💔 The Equity Problem
The training gap isn’t evenly distributed. RAND data shows that teachers and principals in higher-poverty schools are about half as likely to receive AI guidance compared to their wealthier counterparts. The poorest schools are also less likely to use AI tools at all.
This creates a two-tier system: well-resourced districts where teachers get training, students learn to use AI critically, and policies are thoughtful — and under-resourced districts where teachers wing it, students use AI without guidance, and policies either don’t exist or are so restrictive they prevent meaningful engagement.
One policymaker captured the dynamic: “You have policy and what’s actually happening in the classrooms — those are two very different things.”
👩🏫 What Teachers Are Saying
The EdWeek survey captured the frustration directly:
- “I would really like to have some in-depth training on the use of some of the new educational AI tools. Our district has not provided anything at all and it is too expensive to pursue on my own.” — High school math teacher, Colorado
- “I really don’t use it much, but I would be open to some PD on it to learn more.” — High school English teacher, Kansas
- “I need to explore AI before I commit to it.” — High school English teacher, New York
- “I feel that we are at a disadvantage.” — High school social studies teacher, Ohio
Some teachers are finding AI genuinely helpful. A special education teacher in Pennsylvania told EdWeek that ChatGPT was “extremely helpful” for generating IEP goals across 20+ curriculums. A middle school science teacher in Maine used AI for academic advising goal-setting, saying “it saved me an incredible amount of time.”
But these are the exceptions — the teachers who sought out AI on their own. The majority are still in the dark.
🔒 The “Shadow Use” Problem
The 74’s survey also found what one policymaker called “shadow use of AI” — employees implementing generative AI without explicit school or district IT or security approval. When teachers aren’t trained and policies are vague, they use tools anyway. Unvetted. Unguided. Without understanding the privacy implications or the limitations.
That’s not negligence. It’s what happens when you mandate policy without providing training.
🔍 The Bottom Line
Having an AI policy without trained teachers is like having a speed limit without driver’s ed. The policy exists on paper, but the people on the road don’t know the rules. Schools need to stop treating AI training as optional and start treating it as infrastructure. Every dollar spent on policy without matching investment in teacher professional development is a dollar wasted — and the schools that can least afford to fall behind are the ones falling behind fastest.
Sources
- Carnegie Learning: 2025 State of AI in Education Report (74% district policy figure)
- EdWeek Research Center: National survey of 1,135 educators (58% no training figure)
- The 74: Survey of National Association of State Boards of Education members
- RAND: American School District Panel data on AI guidance disparities
- Gallup: February 2025 teacher AI usage poll (60% using AI for work)
- ISTE+ASCD: Commentary on professional development gaps
