Schools across the U.S. are shifting from banning AI chatbots to teaching students how to use them - and how to spot when they're wrong. The change comes as educators try to prepare students for an AI-driven workplace without letting them outsource critical thinking to the tools.
In Charleston County, South Carolina, the state's second-largest school district is building its own AI literacy program after the state failed to publish official guidance. The 50,000-student district hired AI For Education to help draft policy, train staff, and develop courses for middle and high school students. The program teaches students how generative AI works, why it invents information, and how to protect their data when using chatbots.
The district's approach reflects a growing consensus among educators: AI literacy means more than learning to write good prompts. It also means knowing when not to use the technology.
"Good AI literacy," said Rebecca Winthrop, director of the Center for Universal Education at the Brookings Institution, "includes knowing when not to use it."
Teaching students to distrust AI
During a summer training session in Charleston, instructor Amanda Bickerstaff, founder and CEO of AI For Education, prompted an AI tool to "create a map of the world." The result was riddled with errors - Mali spelled "Mail," Egypt labeled "Sopth," and a country called "Africa" in place of Libya. The demonstration is designed to show teachers how quickly chatbots can fail with confidence.
"If you ever want kids not to over trust these tools, try the map demo," Bickerstaff said.
The Charleston curriculum walks students through familiar scenarios, like asking a chatbot to help with a research paper, to show that AI can deliver wildly inaccurate information, including fabricated studies, with unflinching authority. A course module advises: "Always verify any factual information you get from GenAI, especially for your classwork." A section on data privacy warns students that conversations can be stored, used for training, and potentially leaked.
Ray Knauer, a high school teacher who attended the summer training, plans to show examples of AI bias and hallucinations to spark discussion in his AP Research class.
"We're all trying to figure this out right now," Knauer said. "What's the right balance so we don't scare students away from AI but don't encourage them to run to it."
Statewide approaches to AI literacy
Thirty-seven states have published official AI guidance for schools. Utah took a more aggressive approach: in 2024, the state board of education named Matt Winters as its AI education specialist, making it the first state to create a full-time position overseeing AI in schools. Winters has since led training for more than 7,000 teachers, almost a third of Utah's public school instructors. State law requires all districts to have AI policies in place by July 2027.
"There is a clear method from the top in Utah," said Chris Agnew, director of the Generative AI for Education Hub at Stanford University. "When you're tackling big change, it's easier to have coherence if everyone is singing from the same hymn book."
Utah also negotiated data privacy agreements and discounted prices on AI tools, giving rural and under-resourced districts access they might not otherwise have. Maine, West Virginia, and Georgia have since replicated Winters' position.
In other states, AI approaches vary widely from district to district. That inconsistency concerns advocates who worry that students in under-resourced districts won't get the same training. Emma Moss, who oversees AI at the Canyons School District in Utah, led a team that traveled the state delivering over 150 AI literacy trainings funded by a $500,000 grant from a local healthcare provider.
"If I had to give advice: Figure out how to make training accessible to everyone," Moss said. "We said, 'We will come to you.' Whether you are in a rural area or urban, it didn't matter."
Kristina Yamada, the state board's digital technology specialist, said teachers need to change how they think about students using AI to cheat. For younger kids, she recommends conversations like "People can lie. A program can lie. A robot can lie." For older students, she suggests a different framing: "Sure, you can use ChatGPT to write computer code. But when the program stops working, you need to be able to fix it."
Why this matters for educators
Teachers who want to build AI literacy into their classrooms need a curriculum that goes beyond prompt-writing. Start with how AI works, why it produces false information, and when students should not use it. Training students to verify AI output and protect their data is as important as teaching them how to ask questions. For educators seeking structured materials, AI Learning Path for Teachers offers courses specifically designed for classroom use, while AI for Education covers the broader topic of AI literacy in schools. The Charleston and Utah examples show that effective programs are being built on two pillars: teacher training and student instruction that treats AI as a tool with known limitations, not a source of reliable answers.
Your membership also unlocks:
