Artificial Intelligence (AI)I is quickly becoming part of everyday schoolwork. Students might use it to brainstorm ideas, clean up a draft, study for a test, or understand a tricky concept in a new way. But as AI becomes more common, students and teachers are also wrestling with some big questions: When is AI use okay? When should students say they used it? And when does AI stop being a helpful tool and start replacing the student’s own work?

A recent study from Michigan Virtual looked at how students and teachers think about responsible AI use, what guidance they have received, and where there may still be discrepancies in understanding. The findings suggest that AI is already part of many students’ school experiences, but that students and teachers may be clearer about what not to do with AI than about how to use it appropriately to support learning.

In this Interview with a Researcher, we talk through the main takeaways from the study and what they mean for educators.

Why was it important to study how students and teachers think about AI use?

AI is no longer a future issue. It is already here. Many students are using AI tools for schoolwork, whether that means brainstorming, editing, studying, or asking for another explanation of something they do not understand.

At the same time, schools are still figuring out what responsible AI use should look like. Students may hear one message from their face-to-face school, another from an online course, and something else from a teacher, syllabus, or orientation.

So we wanted to understand more than just whether students and teachers know about AI. We wanted to know how they are making sense of it. When do they see AI as a helpful support? When do they think it crosses a line? And where are students and teachers still unsure?

That is especially important in online learning, where students often have to make decisions more independently.

What did you learn about how students are using AI for schoolwork?

One of the clearest findings was that many students are already using AI for school. A majority of students in the study said they had used AI for schoolwork in the past 30 days.

The most common uses were pretty practical. Students reported using AI to brainstorm ideas, edit for grammar or clarity, and explain concepts in a different way. So, for many students, AI seems to be serving as a general support tool. It helps them get started, improve their writing, or better understand something.

But not every student is using AI. A sizable group said they had not used AI recently for schoolwork. And the open-ended responses showed that this was not always just because they had not gotten around to it. Some students deliberately chose not to use AI because of ethical, environmental, personal, or academic concerns.

So the story is not simply “students are all using AI now.” It is more mixed than that. AI use is common, but students are approaching it in different ways and with different levels of comfort.

How did students and teachers compare in their understanding of AI?

Overall, both students and teachers did well on basic AI knowledge questions. Most understood concepts such as AI hallucinations, privacy risks, and algorithmic bias.

Where things got more interesting was when they had to apply their knowledge to a specific schoolwork situation. For example, if an assignment allows AI for brainstorming but not for writing full sentences, what would responsible AI use look like?

Teachers were somewhat more likely than students to answer that kind of question correctly. That suggests students may understand general AI concepts, but applying those ideas to a specific assignment or classroom rule can be harder.

In other words, knowing what AI is is one thing. Knowing exactly how to use it within school expectations is another.

Did students feel like they understood the rules around AI?

In many ways, yes. Most students said they understood which uses of AI were allowed and which were not. Many also said they knew how to use AI responsibly and how to check AI-generated information against reliable sources.

But when we looked more closely, the picture became less clear. Students were much less confident about when and how they should disclose AI use. Only about six in ten students said they knew when and how to disclose AI assistance.

That is an important difference. Students may feel like they understand the big idea—“use AI responsibly”—but still be unsure about the details. For example, Do I need to say I used AI if it only helped me brainstorm? What if it helped me revise one sentence? What if I used it to study but did not submit any AI-generated content?

Those gray areas are where students seem to need more concrete guidance from their teachers or their schools.

Were students and teachers on the same page about what counts as acceptable AI use?

They were more on the same page about what is not okay than about what is okay.

Students and teachers mostly agreed that certain uses crossed a line. For example, they generally agreed that it was not okay for a student to copy an AI-generated answer into an assignment, use AI-generated text without acknowledging it, or submit AI-paraphrased source material as their own writing.

But there was less agreement around support-oriented uses. Teachers were more likely than students to say it was okay for AI to suggest an outline, offer project ideas, summarize readings for studying, or help with grammar and clarity—especially when the student still did the actual thinking and writing.

However, students were more cautious. Even when the AI use sounded supportive rather than dishonest, students were not always sure whether it was acceptable.

That tells us something important: students may have a clearer sense of cheating or misuse than they do of responsible use.

How does this connect to Michigan Virtual’s 2025 statewide AI research?

This study aligns closely with Michigan Virtual’s 2025 AI snapshot report, which found that educators’ use of AI is growing, but trust, comfort, training, and policy clarity are still uneven.

We saw a similar pattern with students. Many students are already using AI, and many say they feel confident about responsible use. But when we look at specific situations—especially disclosure and support-oriented uses—there is still a lot of uncertainty.

So across both studies, the message is similar: AI use is increasing quickly, but shared understanding is still catching up.

It is not enough to give people access to AI tools. Students and educators also need clear expectations, examples, and support so they know how to use those tools well.

Where are students getting guidance about AI use?

Students reported receiving guidance from several different sources. Nearly half said they received guidance from their face-to-face school. Others mentioned Michigan Virtual course orientation, course syllabi, teacher announcements, mentors, peers, and/or school staff.

But nearly one in four students said they had not received any guidance about AI use for schoolwork.

That is a big deal. If students are getting guidance from many different places—or not getting any guidance at all—they may be left to figure out the rules on their own.

This may help explain why students seemed more confident about clearly inappropriate AI use than about more nuanced circumstances. It is easier to know that copying an AI-generated answer is not okay. It is harder to know whether using AI to brainstorm, summarize, or revise is okay if no one has clearly outlined the expectations.

What should educators take away from these findings?

One of the biggest takeaways is that AI guidance should not only tell students what they cannot do. It also needs to show them what responsible use can look like.

Students and teachers seemed to share a pretty strong understanding of misuse. They mostly agreed that copying, misrepresentation, failing to verify information, or failing to acknowledge AI use entirely were wrong.

But there was less clarity around appropriate support. Therefore, students may benefit from understanding very specific examples of appropriate AI use, such as:

“AI can help you brainstorm possible topics, but you need to choose and develop your own idea.”

“AI can give grammar suggestions, but you are responsible for reviewing and revising your own work.”

The more concrete the guidance is, the easier it will be for students to understand the difference between using AI to support their learning and using AI to replace their own work.

What questions remain for future research?

There is still a lot more to learn. An important question remains regarding how students interpret AI guidance once they receive it. Do specific examples help? Do assignment-level instructions make a difference? Does a conversation with a teacher change how confident students feel?

It would also be useful to learn more about students who choose not to use AI. Some students may avoid AI because they do not need it, while others may have serious concerns about ethics, the environment, privacy, or academic integrity. Understanding those perspectives can help schools create guidance that feels balanced and respectful, rather than simply assuming everyone wants to use AI.

As AI becomes more prevalent in education, the goal should not be to simply tell students “use it” or “do not use it.” The bigger goal is to help students and educators build a shared understanding of how AI can be used responsibly, transparently, and in ways that actually support learning.

Overall, this study points to a key challenge for schools and online learning programs: students and teachers may be clearer about AI misuse than they are about responsible AI-supported learning. As AI use continues to grow, students will need more than broad reminders to “use AI responsibly.” They will need clear examples, consistent expectations, and practical guidance about when, how, and why AI can be used appropriately.

Be sure to check out the full research report for more.

AI assisted Michigan Virtual with the creation of this blog post based on original content. The final content was reviewed, revised, and edited by the author.