The volcano cross-section looks incredible. The lava glows, the labels are crisp, and it took your child about ninety seconds to produce with an image generator. Now comes the question that decides whether this project gets full marks or a difficult conversation: is that picture actually allowed? The honest answer is that there is no single answer — and learning to navigate that is now part of doing a school project at all.
Option 1: Not allowed — the school’s integrity rules cover it
Many schools and districts have folded AI into their existing academic integrity policies, and where the policy says student work must be the student’s own, submitting a generated image as project work can fall foul of it — especially if the image is presented as something the student made. The strictness varies enormously. Some schools ban AI content in assessed work completely. Some teachers ban it for their classroom regardless of the district’s position. In this world, the image is simply out, the same as copying a paragraph from an encyclopedia used to be.
Watch for the important distinction these policies usually draw: using AI somewhere in the process (brainstorming, planning, checking) versus submitting AI output as the work itself. Strict schools typically prohibit the second, not the first — but “typically” is doing a lot of work in that sentence, which is exactly why the grey zone in option 4 exists.
Option 2: Allowed, with credit — disclosure is the whole deal
A growing number of schools treat AI images the way they treat photographs from the internet: usable, if you say where they came from. Under this model, the volcano diagram is fine — labeled. A caption such as “Image generated with AI, from my description” turns a potential integrity problem into a citation, the same move students already make for stock photos and charts.
This approach has a real educational logic behind it. Sourcing is a skill schools are trying to teach anyway, and guidance at the international level leans the same way: UNESCO’s guidance on generative AI in education encourages schools to set clear, age-appropriate rules and teach students to be transparent about AI use rather than pretending the tools do not exist.
Option 3: It depends what the image is for
This is the most common real-world answer, and the most sensible. Teachers are usually assessing something specific, and the image’s role decides whether it helps or replaces the learning.
Decoration and illustration: a generated cover image, a background scene for a story, a visual for a poster about volcanoes. The assessed work is the writing, the research, the understanding. Most teachers who allow AI at all allow this.
The work itself: an art project where the drawing is the assignment. A history project where students must interpret a primary-source image. A science fair where the diagram demonstrates the student’s own understanding of the process. Here, generating the image skips the exact skill being graded, and even AI-friendly teachers will usually say no — the same way a calculator is welcome in physics and banned in an arithmetic test.
If you are unsure which bucket a project falls into, that uncertainty is itself the answer: ask. It takes one sentence to a teacher and removes the entire gamble.
Option 4: Nobody has actually said — the grey zone
Plenty of students are working under no clear rule at all: the school has not updated its policy, the teacher has not mentioned AI, and the project is due Friday. The grey zone tempts people to treat silence as permission. It is safer to treat silence as a question. Two habits protect you here. First, disclose anyway — a caption costs nothing, and a teacher who discovers an unlabeled AI image assumes the worst about why it was unlabeled. Second, keep your prompt and process. Being able to show the description you wrote and the choices you made demonstrates that the thinking was yours, even where the rendering was not.
It also helps to know how the adults around this debate are thinking; how teachers really feel about students using AI for homework shows that most are less interested in banning tools than in knowing the work reflects the student.
How the answer shifts with age
The same image means different things at different school levels. In elementary school, projects are about exploration, and many teachers happily allow generated images as long as the child did the learning — some even run the generator in class, because prompting it well is itself a small lesson in describing what you see. The artwork a six-year-old “makes” with AI is judged as play, not authorship.
By high school, the work carries grades that feed transcripts, and expectations tighten accordingly. Assessed coursework in art, design, and media subjects is where the line bites hardest: those portfolios are supposed to evidence the student’s own developing skill, and examiners in several systems now require declarations about how work was produced. A generated image slipped into a portfolio is not a shortcut there; it is misrepresentation of the exact thing being certified. College applicants should assume the strictest version applies to anything submitted as evidence of their own ability.
What a good credit actually looks like
Disclosure only works if it is specific. “Made with AI” is a start, but the strong version says what the student contributed: “Cover image generated from my written description of the water cycle; diagram labels and layout by me.” That one line does three jobs at once — it is honest about the tool, it claims credit for the genuine work (the description, the selection, the editing), and it shows the teacher a student who understands the difference. Students are often surprised to find that teachers respond well to this. The image was never the problem. The hiding was.
A practical habit that makes all of this easy: keep the prompt. Saving the exact words you gave the generator takes two seconds and proves process better than any after-the-fact explanation. If a teacher ever asks, you are not reconstructing your honesty from memory; you are showing it.
One practical note on asking, because students regularly get it wrong in both directions: ask before the work is done, and ask in a form you can keep. A question in the school’s messaging app or an email — “I would like to use an AI-generated image for the cover, credited as generated; the research and writing are mine; is that all right?” — takes a minute, creates a record, and almost always gets a reasonable answer, because teachers would rather be asked than surprised. Asking afterward, or asking in a way designed to be misunderstood (“can I use pictures from the internet?”), converts a policy question into a character question, and character questions are graded much more harshly. The five minutes spent asking properly is the cheapest insurance in this entire article.
The ownership wrinkle most students miss
There is one more catch, and it surprises adults too: an AI-generated image may not legally be yours to claim. Copyright law in the US generally protects human-created work, and the US Copyright Office’s pages on copyright and AI explain that purely AI-generated material may not be copyrightable at all. For a school project, that rarely matters. For a student entering that artwork in a competition, selling it, or using it in a portfolio as their own creation, it can matter a great deal — another reason the credit line is doing more work than it appears to.
Where the stakes jump: fairs, competitions, and portfolios
Everything above gets sharper the moment the project leaves the classroom. Science fairs, art competitions, and design contests judge entries against other students’ work, which changes the ethics completely: an uncredited AI image is no longer just a shortcut past your own learning, it is an advantage taken over everyone who followed the rules. Most organized competitions have now written explicit AI policies — some prohibit generated imagery outright, some require disclosure on the entry form, and ignoring those terms can mean disqualification long after the ribbons are handed out.
Portfolios deserve the same care. A portfolio is a claim: this is what I can do. Colleges and employers read it that way, and a generated piece inside it makes that claim about the wrong person. There is a legitimate version — including AI-assisted work that is labeled as such, in fields like media and design where directors genuinely want to see that fluency — but the label is the whole difference between demonstrating a modern skill and borrowing one. When the work will be judged, compared, or used to win a place, the safest assumption is the strictest one: your own hand, or a disclosure in plain sight.
The verdict
Weigh the four situations and a simple rule covers almost all of them: the image is allowed when it illustrates your work and you credit it; it is not allowed when it replaces your work or hides its origin. Ask when the rule is unstated, disclose when you use it, and never let a generated picture stand in for a skill the project was designed to teach. Students who build that habit now are not just avoiding trouble — they are learning the exact judgment every workplace is about to expect from them.
