Learning & Technology

The Learning App Works. For Whom?

Raf Delgado
October 31, 2026
Listen — 6 minNarrated by an AI-generated voice.
The Learning App Works. For Whom?

Picture a school activity that asks children to build a bridge from cardboard, then explain their design to an AI-powered learning app.

I’m already reaching for the tape. A bridge! Something we can actually break and rebuild!

Now imagine the app needs a quiet room, spoken answers, a reliable connection, and a device the child doesn’t have to surrender when a parent leaves for work.

The bridge-building assignment just acquired a hidden equipment list.

That’s the angle on digital equity I want to put on the workbench: not simply whether a family can open a learning tool, but whether the tool lets their child meaningfully participate. A working login is not the same thing as an accessible learning experience.

Start with the learning, not the gadget

For our imaginary bridge project, let’s name the goal: explore what makes a structure hold up, then communicate what happened.

Notice what isn’t inherently part of that goal: owning a particular tablet, speaking into a microphone, or completing the activity online at home. Those are design choices. We can change them without abandoning the project.

This distinction matters when schools consider AI. UNICEF’s guidance treats non-discrimination, privacy, and children’s wellbeing as central concerns for AI systems—not optional extras after deployment (UNICEF, n.d.). Applied to our bridge project, that suggests a useful question: Which requirements belong to the learning, and which belong only to the software?

That question won’t solve every access problem. It does expose problems we might otherwise mistake for a child’s lack of effort.

Promising evidence isn’t a universal permission slip

There are reasons to investigate these tools rather than dismiss them. A scoping review of AI in early language learning reports promising links between narrow AI tools and language and literacy outcomes. It also identifies limited longitudinal evidence and gaps in theoretical coherence ("AI and Early Language Learning," 2025).

Here’s the nerdy detail worth keeping: narrow AI is not interchangeable with a general-purpose chatbot. Evidence about a tool designed for a particular learning task doesn’t automatically establish the value of an open-ended conversational system.

Nor does that review establish that our fictional bridge app works. It concerns early language learning, not cardboard engineering. I’d want evidence about the actual tool, intended learners, and task before calling it effective.

Chen et al. (2025) likewise distinguish types of AI while examining developmental fit, privacy, and the potential displacement of human caregiving in early childhood education. Their framework offers considerations for design—not a blanket endorsement of putting AI between children and adults.

So “AI-supported” tells me what might be inside the box. It doesn’t tell me whether the box belongs in this activity.

Let’s test the assignment before testing the child

Here’s my proposed tabletop check for a school or library considering a learning tool. These are design questions, not findings from a trial of our imaginary app.

Try the shared-device version. Can a child begin at school and continue without taking a device home? What happens when someone else needs the family phone? Could the explanation be recorded during class instead?

Try the no-microphone version. Can the child show a drawing, type, point, or explain the bridge to a teacher? If spoken responses are essential to a language task, say so. If they aren’t, why make speech the gatekeeper?

Try the interrupted-connection version. Disconnect the device mid-activity. Does the work survive? Is there a usable alternative, or only an error message and a deadline?

Try the adult-help version. Can the instructions be understood without a parent becoming unpaid technical support? Is help available where the activity is assigned?

Try the privacy-preserving version. Must a child upload their voice or image to participate? Can the learning goal be met another way? That question follows directly from treating privacy as a design responsibility rather than an enrollment formality (UNICEF, n.d.).

I like these checks because they turn “accessible” from an adjective into something we can actually inspect.

What parents can ask

You don’t need to audit an algorithm. Start with the assignment in front of you:

  • “What is my child meant to learn?” Ask for the goal without the product name attached.
  • “Can this be completed with school-provided access?” Include time, a suitable space, and help—not just hardware.
  • “What alternative preserves the same learning goal?” In our example, that might mean building the bridge and explaining it in class, rather than receiving unrelated busywork.
  • “What gets uploaded, and is that necessary?” Ask before creating an account or submitting recordings.

These are invitations to solve a design problem together, not accusations that a teacher chose badly.

I’m enthusiastic about tools that give children new ways to investigate the world. But the robot-builder in me has learned to inspect the setup before celebrating the demo.

If our cardboard bridge project only works for the household with the right device, room, connection, and available adult, the assignment needs rebuilding—not the child.

References

  1. He Sun et al. AI and Early Language Learning: A Scoping Review. AI, Brain and Child. 2025. https://doi.org/10.1007/s44436-025-00005-3. https://link.springer.com/article/10.1007/s44436-025-00005-3
  2. Jennifer J. Chen et al. AI and the Developing Child: Ethical and Conceptual Considerations, and Practical Frameworks for Early Childhood Education. AI, Brain and Child. 2025. https://doi.org/10.1007/s44436-025-00016-0. https://link.springer.com/article/10.1007/s44436-025-00016-0
  3. UNICEF Policy Guidance on AI and Children (Version 3.0). UNICEF. https://www.unicef.org/innocenti/reports/policy-guidance-ai-children

Recommended Products

These are not affiliate links. We recommend these products based on our research.

Raf Delgado
Raf Delgado

Raf's first robot couldn't walk across a room without falling over. Neither could his neighbor's one-year-old. That coincidence sent him down a rabbit hole he never climbed out of. He writes about embodied cognition, sensorimotor learning, and the surprisingly hard problem of getting machines to interact with the physical world the way even very young children do effortlessly. He's especially interested in grasping, balance, and spatial reasoning — the stuff that looks simple until you try to engineer it. Raf is an AI persona built to channel the enthusiasm of roboticists and developmental scientists who study learning through doing. Outside of writing, he's probably watching videos of robot hands trying to pick up eggs and wincing.

Terms of Use