AI in Schools

AI in schools refers to the use of artificial intelligence tools by students, teachers and school systems. Students use AI chatbots for homework help and research. Teachers use AI to plan lessons, create materials and handle paperwork. Some schools use AI tutoring and learning software. Use is spreading faster than school policies, and the research on how AI affects learning is still early and mixed.

How AI is being used in schools today

AI arrived in classrooms mostly from the bottom up: students and teachers started using general tools like ChatGPT before most schools had rules for them. U.S. survey data shows how quickly that happened.

How students use AI

• Homework and schoolwork help. In a Pew Research Center survey conducted in fall 2025, just over half of U.S. teens said they had used AI chatbots for help with schoolwork. One in 10 said they do all or most of their schoolwork with a chatbot's help.
• Rapid growth. In RAND's American Youth Panel, the share of students in middle school and up who used AI for homework help rose from 48% in May 2025 to 62% in December 2025.
• Common tasks. Typical student uses include explaining concepts, researching topics, getting feedback on writing, practicing problems and, in some cases, producing work they submit as their own.

How teachers use AI

• Widespread use. In a spring 2025 Gallup survey with the Walton Family Foundation, six in 10 U.S. public K-12 teachers said they had used an AI tool for their work during the school year.
• Time savings. Teachers who used AI at least weekly estimated saving 5.9 hours a week, which Gallup calculated as about six weeks over a school year. That figure is self-reported and applies only to weekly users.
• Common tasks. Typical teacher uses include drafting lesson plans, creating quizzes and worksheets, adapting materials for different reading levels, writing parent communications and giving feedback.

How schools and districts use AI

Some schools adopt AI-powered tutoring platforms, learning software that adjusts to each student's level, and tools for administrative tasks. Districts are also training staff: RAND found that roughly half of U.S. school districts reported training teachers on AI by fall 2024, twice as many as the previous fall.

The rules: school policies and government guidance

School policies have not kept pace with use.

• Many schools lack clear policies. RAND surveys from the 2024 to 2025 school year found that only 45% of principals reported having school or district policies or guidance on AI. More than 80% of students said their teachers had not explicitly taught them how to use AI for schoolwork.
• Most teachers lack formal guidance. A Gallup survey conducted in early 2026 found that just 18% of U.S. K-12 teachers receive formal guidance on how to use AI at work, and about a third receive none.
• States are issuing guidance. As of August 2026, a tracker maintained by the education training organization AI for Education counts more than 30 states with official guidance or policy on AI in K-12 schools. Other trackers count differently, because definitions of "guidance" vary.

At the federal level in the U.S., an April 2025 executive order set a policy of promoting AI literacy and educator training. In July 2025, the U.S. Department of Education issued guidance saying federal grant funds can support uses such as AI-enhanced tutoring, and stressed engaging parents in decisions about new technologies. In April 2026, the Department finalized a grant priority for AI in education, which applies to certain competitive grants and does not require schools to do anything.

Internationally, UNESCO's 2023 guidance on generative AI in education urged governments to regulate these tools, including data protection standards and a recommended age limit of 13 for using AI tools in the classroom.

Privacy laws also apply. In the U.S., the Family Educational Rights and Privacy Act (FERPA) gives parents rights over their children's education records at schools that receive U.S. Department of Education funding, with those rights passing to students at 18 or when they enter college, and the Children's Online Privacy Protection Act (COPPA) sets rules for collecting personal information from children under 13. The Federal Trade Commission has said schools can consent on parents' behalf only when student information is used for a school-authorized educational purpose and no other commercial purpose.

Consumer AI tools also have their own age rules. OpenAI's terms of use, for example, require users to be at least 13, or the minimum age required in their country, and users under 18 need a parent's or guardian's permission.

What the research says about AI and learning

Research on AI and learning is new, and most studies are small or short. So far, how AI is designed and used appears to matter more than whether it is used at all.

Well-designed AI tutoring has shown strong results in some studies. In a randomized controlled trial published in Scientific Reports in 2025, Harvard physics students who learned from a custom AI tutor built on established teaching practices achieved median learning gains more than double those of students in an in-class active learning session, in less time. The study covered college students in one course, and the authors cautioned against using it as a reason to replace in-class teaching.

Structured programs can help in lower-resource settings. A 2025 World Bank working paper reported that secondary students in Nigeria who used an AI chatbot for English tutoring in a six-week, teacher-guided after-school program improved significantly on an assessment covering English, AI knowledge and digital skills. The study is a working paper and has not been peer-reviewed.

Unrestricted AI can undermine learning. In a field experiment with nearly 1,000 high school math students in Turkey, published in PNAS in 2025, students given a standard AI chatbot did better on practice problems while they had it. Once access was removed, they scored 17% lower than students who never had it. A version designed with learning safeguards largely avoided that harm. The study covered math in one country.

Getting the task done is not the same as learning. The OECD's Digital Education Outlook 2026 concluded that successfully completing tasks with generative AI does not necessarily translate into learning, and that tools designed for teaching and learning show more lasting benefits.

The practical takeaway from the evidence so far: AI designed to guide students' thinking has helped in some studies, while AI that simply supplies answers can let students skip the thinking that learning requires.

AI in schools: pros and cons

The lists below mix research findings with concerns that are still being studied. Where evidence exists, it is noted.

Potential benefits of AI in education

• Personalized help. AI tutors can explain concepts in different ways, at a student's pace and at any hour. Well-designed tutors have produced strong learning gains in early trials.
• Time back for teachers. Weekly users report saving hours on planning, materials and paperwork.
• Differentiated materials. Teachers can quickly adapt readings and exercises for different levels and needs.
• Instant feedback. Students can get immediate responses on practice work and drafts rather than waiting days.
• AI literacy. Using AI with guidance helps students learn how these tools work, where they fail and how to use them responsibly.

Risks and concerns

• Shortcutting learning. When AI does the work, students may not build the underlying skills. Research on unrestricted chatbot use supports this concern.
• Academic integrity. In Pew's fall 2025 survey, 59% of teens said using AI to cheat is a regular occurrence at their school. That reflects perception, not measured cheating rates.
• Inaccurate information. AI chatbots can produce confident, false answers. Students need to know how to check them. What Is an AI Hallucination? explains why this happens.
• Unfair accusations. AI detectors can wrongly flag student writing. A 2023 Stanford-led study found that seven popular detectors misclassified, on average, more than half of 91 essays by non-native English writers as AI-generated.
• Privacy. Student data entered into AI tools may be stored or used in ways families do not expect.
• Critical thinking concerns. In RAND's December 2025 survey, 67% of students agreed that the more students use AI for schoolwork, the more it will harm their critical thinking skills.
• Unequal access and training. Access to good tools and trained teachers varies between schools and districts.

Can teachers tell if students use AI?

Not reliably. Teachers often notice writing that does not match a student's usual work, invented sources or answers that miss the specifics of an assignment, but AI detection tools are not accurate enough to prove AI use. OpenAI's own guidance for educators says its research did not find detectors reliable enough, given that educators could be making judgments about students with potentially lasting consequences.

Many teachers focus on the process instead: in-class writing, drafts and version history, and asking students to explain their work. For a full look at the clues and the limits of detectors, see How to Tell If Something Was Written by AI.

Should AI be used in schools?

This is a question of values and priorities as much as evidence, and reasonable people disagree.

Arguments for include preparing students for a world where AI is common, giving every student access to individualized help, reducing teacher workload and teaching students to use AI critically rather than secretly.

Arguments against include the risk that students skip the struggle that builds skills, concerns about privacy and data use, the cost and unevenness of implementation, and limited long-term evidence on learning outcomes.

A common position in education guidance is a middle one: allow AI for specific purposes, keep teachers in charge, set clear rules and teach students how to use it well. The U.S. Department of Education's 2023 report on AI in teaching and learning, issued under the previous administration, made keeping "humans in the loop" its first recommendation.

Questions parents can ask their child's school

• What is the school's AI policy? When is AI allowed, and when is it not?
• Which AI tools are approved? Has the school reviewed them for student privacy?
• What student data do these tools collect, and who can access it?
• How are students expected to disclose AI use?
• How does the school decide whether a student misused AI? Does it rely on detection software?
• Are students taught how AI works and how to check its answers?

At home, it helps to ask your child how they use AI, and to talk about the difference between using it to understand something and using it to avoid doing the work.

Practical guidance for teachers

• Set expectations assignment by assignment. Students need to know whether AI is allowed for brainstorming, feedback, drafting or nothing at all.
• Design for process. In-class work and oral explanations show thinking that is hard to outsource to AI, and drafts with version history add useful context.
• Verify AI-generated materials. Check facts, reading levels and examples before using AI-created content with students.
• Protect student information. Avoid entering identifiable student information into tools your school has not approved.
• Teach verification. Show students how to check AI answers against reliable sources.