
Image/caption credit: People Magazine
This story about a school using generative AI to create an image of what a student might look like two decades in the future has gone viral. It started with an Instagram post and has since been covered by People and other outlets. I believe it is a very important cautionary tale.
“What do you want to be when you grow up?”
Schools often ask students to imagine what they want to do/want to be when they grow up. Kids will usually come up with answers using what they know. For example, if they have family members who are police officers and firefighters, that’s what they may say they want to do, and if they know all about princesses and dragon trainers, they might say that’s what they want to be.
This type of activity is a seemingly benign way for teachers to gain a little bit of insight into a student’s interests, their schema outside of school, and even the scope of their imagination. Some teachers may even believe that for students who struggle in school or with motivation on school-related tasks, the idea of working toward a dream job or future career can help them with those tasks. Whatever the reason teachers have for using an activity like this, students should never be made to feel like whatever they decide at any moment is what they have to do or become or that they don’t have other options.
In the case of this viral story, one little girl, the six-year-old daughter of writer Nikki Anderson, thought she might like to be a nurse and created a construction paper picture to illustrate that idea. It was cute. Then, her teacher uploaded her image into a generative AI tool and produced a photorealistic image of “her” in the future as a 25-year-old nurse.
Except, it’s not her.
And it will never be her, not her exact hair, her exact complexion, her exact body type. It’s a disturbing realistic-looking AI-generated mash up of other people, of other images.
Anderson is quoted in the People article as saying, “Teachers are already being asked to do an extraordinary amount with limited resources, and I don’t think the responsibility for navigating an entirely new tech landscape should fall exclusively on them.”
And, I agree.
Expecting teachers (even expecting school and district administrators) to figure out safe ways to use generative AI, which by its own design is intentionally and extremely volatile, is unfair and dangerous.
There are so many things to further unpack from this one story:
-
- The psychological impact that seeing images “of themselves” like this can have on young children.
- Student data and privacy concerns when teachers and school districts use AI. (Some of which are outlined by Anderson in this follow up Instagram post.)
- The fact that permission granted by parents to publish an image of a child (in a newsletter or district website) is not the same as permission to upload and manipulate those images in generative AI tools.
- AI bias and how it is influenced by and perpetuates societal expectations of what is considered successful, attractive, acceptable, etc.
- How even when people say they use generative AI to support thinking and learning it is actually creating its own material and inauthentic experiences that replace thinking instead.
Parents, teachers and other educators, please…don’t upload pictures of children into AI…don’t limit their imaginations…don’t create impossible ideals for children to compare themselves to — not now, in the future, or ever.
Please don’t.