In American classrooms, something has changed. It’s not strong or loud. There is, however, a quiet tension growing in almost every school district in the country. This tension exists between students who already know how to use AI tools and teachers who are still trying to catch up, as well as between schools that have made policy documents and schools that haven’t touched the issue at all. Researchers at Stanford have known about this gap for a while now, and some of them have been working to close it.
Victor R. Lee, who is a faculty affiliate at Stanford HAI and runs the AI+Education program for the Stanford Accelerator for Learning, has been watching this with a mix of worry and a sense of how important it is. “Kids are racing ahead in their grasp of the technology, and schools are scrambling to catch up,” he’s said. In thousands of school districts right now, that one sentence sums up what’s going on: a slow response from the school to a technology that students have already incorporated into their daily lives.
The situation is complicated by more than just how quickly AI is being used. It’s how the response is all over the place. Some schools have written AI policies, but they’re pretty general and only care about following the rules. They leave most of the real decisions up to the teachers. Some people haven’t written anything at all. In classrooms across the country, students sign disclosures about “responsible use of AI” even though they don’t all agree on what that means in real life. A policy that says “don’t cheat” is not the same as one that thinks about when AI can help you learn and when it can speed you up.
Stanford tries to give teachers real tools instead of just general advice through programs like CRAFT, which was created by Stanford researchers and high school teachers working together. Students can find more and more free resources on CRAFT’s website that help them think critically about AI in any subject. Most school districts can’t make that kind of basic, ready-to-use material on their own. Maybe many schools would just keep winging it through each semester if they didn’t get this kind of outside help.
Take a moment to think about what teachers are really going through. A study published in Springer Nature in 2025 found that teachers are really worried about false information, academic honesty, and not being sure enough about new tools to teach their students how to use them correctly. The problem isn’t that teachers don’t want to do it. It’s because progress in this area of work has been slow, uneven, and sometimes not even happening. Because they’re not built for speed, seeing an institution try to deal with something moving this quickly is awkward.
Lee’s advice for parents talks about the social side of things, which is something that most policy documents don’t. He’s not just worried that students might hire ChatGPT to do their essays for them. When AI is added to phones and social media, it could make young people even more alone and dependent on screens at a time when face-to-face interactions are shaping who they become. “With mobile phones, video games, and social media, we’re getting in the habit of being alone, and AI can further that,” he notes. There are important things that that quiet observation means.
It seems like the schools that get this right in the next two or three years will set standards and habits that will last for a whole generation. The work that Stanford does isn’t a magic bullet. But it might be one of the more serious efforts to bring the country’s conversation together into something that makes sense. There is still no clear answer to the question of whether school districts will act quickly enough to use it.
