AI Images of Students: Consent Lessons From a 'Future Self' Photo

Harper, 6, came home from school in October 2026 with two pictures of her future. One was a construction-paper drawing she made herself. The other was an AI-generated image of her at 25, dressed as a nurse. Her mother, Nikki Anderson, hadn’t been asked about the second one. “Did a teacher really upload my daughter’s photo somewhere to make this, knowing what we know about AI?” she told People. The activity was almost certainly well-meant. But it raises the consent questions about AI images of students that most schools haven’t answered yet: who approved the tool, where the photo went, and whether a computer should be doing a 6-year-old’s imagining for her.

Key Takeaways

• A South Carolina school created an AI “future self” image of a 6-year-old without parental consent; the OECD’s AI Incidents Monitor logged it as an AI “hazard,” meaning potential harm, not confirmed harm.
• Under FERPA, a photo that focuses on a specific student and is kept by the school can be an education record, which brings privacy obligations.
• The FTC says schools can consent on parents’ behalf under COPPA only for educational use, and recommends that districts, not individual teachers, decide which online services are appropriate.
• AI image generators carry bias: a 2023 Bloomberg analysis of Stable Diffusion found it over-represented lighter skin in high-paying jobs and darker skin in low-paying ones.
• The mother’s second objection matters too: “Why are we outsourcing” a child’s imagination of her future?

What happened in South Carolina

The case surfaced in October 2026 after Anderson posted about it on social media. People interviewed her for an exclusive report by staff editor Georgia Slater. The school is not named in the coverage.

Anderson’s concerns came in two parts. The first was privacy: “What software was used? Was it an outside platform? What happens to those pictures after? Did any of us actually consent to that?” She added: “Children cannot meaningfully consent to this sort of use; I think our children deserve protection, not only of their image, but of their minds.”

The second was about childhood itself. “6-year-olds are already creative. They’re already innovative. They’re already capable of imagining futures that adults couldn’t possibly predict. Why are we outsourcing that?” The Economic Times, which also covered the case, quoted her asking: “Can we at least leave their imaginations alone?”

She was careful about blame. “I don’t want this to become a story about attacking educators,” she said. “Teachers are already being asked to do an extraordinary amount with limited resources.” At the time of the reports she was waiting to hear back from the principal and school board. Reactions to her post split: some parents were horrified, others thought she was overreacting to a harmless activity.

The OECD AI Incidents Monitor, which tracks AI-related events worldwide, logged the case as an AI “hazard” rather than an “incident,” because no actual harm or rights violation had been confirmed and the school’s response was pending. That’s a fair label. Nothing bad is known to have happened to Harper’s photo. The issue is that nobody had asked.

How an AI “future self” image is made

Understanding the mechanics clarifies the privacy question.

Most “age me up” tools work in one of two ways. Some take the uploaded face, map its features, and have an image model generate an older version that keeps those features. Others take a text description plus the photo as a reference. Either way, the child’s photo is uploaded to whatever service runs the model. That service’s terms decide whether the photo is stored, for how long, who can see it and whether it can be used to improve the company’s models.

A face photo is not ordinary data. It can identify a child, it’s hard to take back once copied, and in many systems it can be turned into a biometric template. A teacher clicking “upload” in a free web app may be agreeing, on behalf of 25 children, to terms nobody at the school has read.

There’s also what the image contains. Generative image models learn from huge collections of online pictures, and they reproduce the patterns in them. A 2023 Bloomberg analysis of more than 5,000 images from the model Stable Diffusion found that it generated people with lighter skin more often for high-paying jobs and over-represented darker skin in low-paying ones; women were underrepresented in high-paying jobs and overrepresented in low-paying ones. Models change, but the lesson holds: an AI picture of “your future” carries the model’s assumptions about who becomes what.

Harper’s own reaction hints at this. According to her mother, she loved the picture “because she had straight and wavy hair and that is what she always wanted, and she hopes she would be that beautiful when she grows up.” A 6-year-old being shown a machine’s idea of her adult beauty is worth a pause, whatever the intent.

What US privacy rules say about AI images of students

There’s no law written specifically for AI images of students yet. But two federal frameworks apply, and both point to the same answer: the school, not the individual teacher, should decide.

Question	What the rules say	Source
Is a student’s photo protected?	A photo or video “directly related to a student” and maintained by the school is an education record under FERPA	US Dept. of Education, FERPA photo FAQ
What makes it “directly related”?	Among other factors, when the photo intends to make a specific student the focus, such as an ID photo	US Dept. of Education, FERPA photo FAQ
Can a school consent instead of parents?	Under COPPA, for children under 13, yes, but only “for the use and benefit of the school, and for no other commercial purpose”	FTC COPPA FAQ
Who should approve the tool?	”We recommend that schools or school districts decide whether a particular site’s or service’s information practices are appropriate, rather than delegating that decision to the teacher”	FTC COPPA FAQ
What if the company uses data for its own purposes?	The operator “will need to obtain parental consent”	FTC COPPA FAQ

The US Department of Education’s FAQ on photos and videos under FERPA explains that a photo becomes an education record when it is directly related to a student and maintained by the school or someone acting for it. A headshot uploaded to make an individual AI portrait fits the “specific student the focus” factor. Whether the AI company “maintains” it on the school’s behalf depends on the contract, which is exactly why a contract should exist.

The FTC’s COPPA guidance is direct. Schools may act as the parent’s agent and consent to an online service collecting a child’s information, but “the school’s ability to consent for the parent is limited to the educational context.” And the agency’s best-practice advice is that districts, not teachers, decide which services meet that bar.

So the honest answer to “was this illegal?” is: we don’t know, and it depends on the tool and its terms. The clearer answer to “should a teacher pick an AI photo app on her own and upload 25 children’s faces?” is no, according to the FTC’s own guidance.

The second question: who gets to imagine a child’s future?

Anderson’s point about imagination deserves as much attention as the privacy one.

Harper made a drawing of her future. Then the AI made a better-looking, more “finished” version, and she was more interested in the AI one. That’s not a moral failing in a 6-year-old; polished images are compelling. But it’s a small, real example of a pattern we’ve written about across this series: when a machine produces the impressive version, the child’s own version starts to look like the rough draft.

There’s a simple design fix that keeps the activity and keeps the child as the author. Let the child imagine, draw or describe their future, and if AI is used at all, have it illustrate the child’s own description, with no photo of the child. We explore that setup in our piece on kids writing while AI only illustrates.

What to do if this happens at your child’s school

Ask four specific questions, in writing

Keep it calm and concrete. “Which tool was used? Who approved it? Were students’ photos uploaded, and if so, where are they stored now? Can they be deleted?” Specific questions get specific answers; vague complaints get vague reassurance.

Ask for deletion if photos were uploaded

You can ask the school to request deletion from the vendor and confirm when it’s done. If the school has a data privacy agreement with the vendor, it should include a deletion process.

Ask whether the district has an approved-tools list

Many districts maintain a list of vetted apps with signed privacy agreements. If the tool wasn’t on it, the fix is a process change: teachers use only approved tools, and new tools go through review. That’s the FTC’s recommendation, and it protects teachers too.

Talk to your child about the picture

If your child brought one home, ask what they thought. Ask which future they like better, the one they drew or the one the computer made, and why. You can mention that the computer was guessing, and that its guesses come from pictures of other people. Our guide to AI image generators, kids and consent has more on these conversations.

What not to do: post your child’s photo or the AI image while complaining

It’s tempting to share the image to make the point. But posting a child’s face, or an AI-aged version of it, spreads exactly the data you’re worried about. Anderson’s case went viral; you don’t need yours to. Our article on sharenting and kids’ privacy explains why.

What to Watch For Over the Next 3 Months

• Week 4: Look at what comes home in backpacks and school apps. Ask your child’s teacher whether any upcoming projects use AI with student photos, voices or names.
• Month 2 red flags: AI-made images or videos of your child appearing without notice, a vendor you’ve never heard of in school communications, or no answer to a written question about where photos are stored.
• Month 3 self-check: Does your district publish its approved-tools list? Does your school’s photo consent form mention AI? If not, it’s a reasonable request to make at a parent meeting.

Frequently Asked Questions

Can a school make AI images of my child without asking me?

It depends on the tool, its terms and your school’s policies. Under COPPA, schools can consent for parents only for educational use with no other commercial purpose, and the FTC recommends districts, not teachers, approve services. Under FERPA, a photo focused on a specific student and kept by the school can be an education record. Ask your school which rules it followed.

What happened with the South Carolina “future self” photo?

In October 2026, a South Carolina mother, Nikki Anderson, found that her 6-year-old’s school had made an AI image of her daughter as a 25-year-old nurse without parental consent. She raised privacy and imagination concerns and was awaiting answers from the school. The OECD’s AI Incidents Monitor logged it as a potential hazard.

Are AI-generated photos of kids a privacy risk?

They can be. Making one usually means uploading the child’s face to an outside service whose terms determine storage, sharing and possible use for model improvement. Face photos are identifying and hard to retrieve once copied. The risk depends on the specific tool and contract, which is why a district review matters.

Is there anything wrong with the image itself?

Possibly. Image generators reproduce patterns from their training data, including stereotypes. A 2023 Bloomberg analysis found Stable Diffusion over-represented lighter skin in high-paying jobs and darker skin in low-paying ones. An AI “future you” also replaces the child’s own imagining with a machine’s guess.

What should I ask my child’s school?

Ask which tool was used, who approved it, whether student photos were uploaded and where they’re stored, and whether they can be deleted. Ask whether the district keeps an approved-tools list and whether photo consent forms mention AI. Keep the tone collaborative; most teachers mean well and lack guidance.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Slater, G. (2026, October). “Mom Calls Out School for Making an AI-Generated Photo of 6-Year-Old Daughter as an Adult (Exclusive).” People. https://people.com/mom-calls-out-school-for-making-ai-generated-photo-of-6-year-old-daughter-as-adult-12184403
2. OECD.AI. (2026, October 20). “South Carolina School’s AI-Generated ‘Future Self’ Photo of Child Sparks Privacy Outcry.” OECD AI Incidents and Hazards Monitor. https://oecd.ai/en/incidents/2026-10-20-3a14
3. The Economic Times. (2026, October). “South Carolina mother ‘grossed out’ after finding an AI-generated photo of what her 6-year-old daughter might look like after 19 years.” https://economictimes.indiatimes.com/world/us/south-carolina-mother-grossed-out-after-finding-an-ai-generated-photo-of-what-her-6-year-old-daughter-might-look-like-after-19-years-asks-can-we-at-least-leave-their-imaginations-alone/articleshow/135085378.cms
4. US Department of Education, Student Privacy Policy Office. “FAQs on Photos and Videos under FERPA.” https://studentprivacy.ed.gov/faq/faqs-photos-and-videos-under-ferpa
5. Federal Trade Commission. “Complying with COPPA: Frequently Asked Questions.” https://www.ftc.gov/business-guidance/resources/complying-coppa-frequently-asked-questions
6. Bloomberg. (2023, June). “Generative AI Takes Stereotypes and Bias From Bad to Worse.” Bloomberg Graphics. https://www.bloomberg.com/graphics/2023-generative-ai-bias/