One of the biggest arguments I hear against artificial intelligence in schools is that students will use it to cheat. When a student can create an essay or solve a math problem in seconds, how do we know whether learning took place?
But did AI create a cheating problem, or expose one that education has faced for decades?
A 2026 study of 4,354 students across six high schools found that approximately 72% reported at least one academically dishonest behavior in the previous month. Researchers described overall cheating as broadly consistent with historical patterns, although their sample was not nationally representative.
Students were cheating long before ChatGPT. AI changes the tools available, but we should not mistake the tool for the entire problem.
Meanwhile, a nationally representative Common Sense Media survey found that 70% of American teenagers use AI for schoolwork, yet only 27% said a teacher had discussed how to use it safely. A recent Fortune article examines the Gates Foundation’s $400 million commitment to AI in education and educators’ concern that technology could widen learning gaps if schools lack the support to use it well.
We are telling students what they cannot do with AI without adequately teaching them what they should do with it.
I am not advocating unrestricted access. I believe that AI education should be intentional, age-appropriate, and balanced. Students need to understand both the risks that come with these tools and the opportunities they can create when they use it thoughtfully.
That means teaching students what to watch out for:
- Accuracy and misinformation: AI can deliver persuasive but incorrect answers. Students must verify claims, examine sources, and question confident-sounding responses.
- Privacy and digital safety: Children need to know why they should not enter personal details, photographs, or other information into tools they do not understand.
- Bias and fairness: Students should ask which perspectives show up in an answer, which don’t appear, and how stereotypes are represented.
- Academic integrity and ownership: AI can help explain an idea, but it should not silently do a student’s thinking. Students need clear expectations for using AI for assistance.
- Human connection and independence: A chatbot is not a substitute for friendships, trusted adults, character development, or problem-solving.
- Environmental responsibility: Data centers use electricity, infrastructure, and sometimes substantial amounts of water for cooling. The International Energy Agency estimated that data centers consumed about 1.5% of global electricity in 2024 and projected that demand would more than double by 2030. Students deserve to know that their digital choices have physical costs.
These are reasons to teach AI, not to avoid the conversation. And there are genuine possibilities worth exploring:
- Personalized learning: Students can request another explanation or practice a challenging concept at their own pace.
- Creativity and innovation: Students can brainstorm, prototype, code, and test ideas they might not otherwise have exposure to.
- Accessibility and inclusion: Various selected tools can support reading, writing, and different ways of showing understanding.
- Research and inquiry: AI can help students with their questions and identify how to investigate, provided they check the evidence themselves.
- Communication and collaboration: Students can practice languages, improve drafts, and consider different perspectives before discussing ideas with real people.
- Environmental problem-solving: Students can explore how AI supports wildlife monitoring, climate research, and more efficient energy use while weighing those benefits against its environmental footprint.
At The Innovation School, our approach centers on projects, passions, peers, and play. Technology can support all four, but the student must remain the thinker and creator. Especially in a K–8 setting, I want children building, talking, reading, playing, and struggling productively. AI should extend those experiences when appropriate, never becoming the default experience itself.
That is also why the concern about widening learning gaps matters. AI may offer new ways to personalize learning, but those benefits depend on students already having strong foundational skills and teachers having the time and training to use the tools well. An AI tutor cannot replace hands-on number sense for a kindergartener or the explicit instruction a child needs to learn to read. An eighth grader needs enough subject knowledge to recognize when an AI explanation is wrong. Access to technology is not the same thing as access to a quality education.
Used thoughtfully, though, AI can support learning without replacing it. A student might ask AI about the cardiovascular system, then build a model, explain how body systems interact, and answer questions. The tool supported the process, but the student demonstrated the learning.
We cannot know what AI will look like when today’s kindergartners graduate. We can teach them to question, create, protect themselves, consider consequences, and recognize when to put technology aside.
The conversation is not whether students should use AI. It is how we help them discover what they can do with and without it.
