Is using AI cheating?

What counts as cheating in an AI-enabled world, and are our children becoming more capable, or more dependent? This leads to the question on everyone´s minds, is using AI cheating?

A student asks AI to explain a difficult concept. Another asks it to write an essay and submits the result as their own. Both have used AI. But they have made very different decisions about learning and honesty.

For parents, teachers, and students, that difference can be difficult to navigate. When does help become a shortcut? How much assistance is too much? And if AI is becoming part of working life, why would using it at school be wrong?

Using AI is not automatically cheating. It can become cheating when students use assistance that is prohibited or misrepresent what they have done themselves. But there is another question we need to ask: even when AI use is allowed, is it helping the student learn?

What does the research tell us?

In an October 2023 Stanford Graduate School of Education article, researchers challenged the assumption that ChatGPT had immediately caused more students to cheat. Their early survey findings showed broadly similar levels of self-reported cheating before and after its release. They also highlighted pressures that existed long before AI: overwhelming workloads, fear of failure, and assignments students found difficult to connect with.

A follow-up published in February 2026 examined responses from 4,354 students across six U.S. high schools. Overall self-reported cheating remained broadly comparable with earlier findings, while AI use increased for both support and completing entire assignments.

There are important limits. The follow-up used surveys collected in 2024, relied on students’ own reports, and compared different groups across studies. It cannot establish how every school is experiencing AI today, or prove that AI has no effect on dishonesty.

More recent evidence shows why the concern deserves attention. Pew Research Center’s February 2026 report, based on a fall 2025 survey, found that 59% of U.S. teens ages 13–17 believed students at their school used AI to cheat at least somewhat often. That measures perceived behavior around them, rather than an admission of their own cheating.

The evidence gives us reasons to take AI misuse seriously while being careful about claims of a universal cheating epidemic.

The same tool can serve different purposes

Consider a 16-year-old preparing an argument about a community issue. They might ask AI to identify weaknesses in their reasoning, then investigate those weaknesses and revise their argument. Alternatively, they might ask AI to produce the entire argument and present it as independently written.

In the first scenario, the learner still has substantial thinking to do. In the second, the finished work may conceal how little they understand.

However, even brainstorming or feedback can be inappropriate if the assignment requires independent work. California’s Department of Education guidance emphasizes clear expectations for each assignment, transparent disclosure, and the distinction between permitted assistance and unauthorized use.

Acknowledging AI use does not make prohibited assistance acceptable. Students need to know the boundaries before they begin, including which tools and functions are allowed. If the instructions are unclear, asking the teacher is part of acting responsibly.

A better answer is not always better learning

Academic honesty matters. So does what happens to the learner’s understanding.

A randomized study published in PNAS in 2025, involving nearly 1,000 high school mathematics students in Turkey, illustrates the difference. Students using a general GPT-4 interface performed better during assisted practice but worse on a subsequent assessment without AI than students who practiced without it. A specially designed tutor with teacher-informed guardrails largely mitigated that harm.

This was one study in one subject and setting. It does not show that every use of AI weakens learning. It does show why a polished answer is insufficient evidence of understanding.

For a parent or teacher, useful questions extend beyond “Did you use AI?” Ask: “What did it help you understand? What did you question or change? Can you explain your reasoning without it?”

What does this mean at Young Steward?

Young Steward’s Fellows program is being developed for learners ages 16–17 around a connected goal: strengthening human innovation alongside responsible AI capability.

Our approach is to make AI use purposeful, transparent, and accountable. Learners need clear guidance on when AI can support their work and when they must work independently.

We also help them understand that expectations may differ between Young Steward and their own school. Through ongoing conversations, we encourage learners to reflect on their choices, ask questions when boundaries are unclear, and follow their school’s rules whenever they use AI for schoolwork.

In project work, the learning should be visible beyond the final presentation. Learners should be able to explain the problem they investigated, the choices they made, where AI contributed, and how they checked and tested its suggestions. Keeping drafts, recording meaningful AI assistance, and discussing decisions with a Steward Guide can help make that process clear.

Imagine a Fellow developing a prototype to address a school challenge. AI might help generate options or produce an early version. The learner’s responsibility is to investigate whether the solution meets people’s needs, identify weaknesses, test assumptions, and explain what still requires human judgment.

That creates an opportunity to teach integrity through practice. Learners can experience both the usefulness of AI and the responsibility that comes with using it.

Our aim is for Fellows to leave with more than an impressive project. They should leave better able to think independently, explain their contribution honestly, and judge when AI assistance helps and when the learning requires them to do the work themselves.

Become a Fellow

References:

• Stanford Graduate School of Education article -"What do AI chatbots really mean for students and cheating?": https://ed.stanford.edu/news/what-do-ai-chatbots-really-mean-students-and-cheating?utm

• PEW Research Center- "How Teens Use and View AI": https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/?utm

• PNAS Study Generative AI without guardrails can harm learning: Evidence from high school mathematics: https://www.pnas.org/doi/10.1073/pnas.2422633122