The more important question is whether they understand how to work with it. Knowing how to type a question into an AI system is not the same as understanding how to communicate a problem clearly, evaluate the response, recognize uncertainty, or decide when an answer should not be trusted.
That distinction has shaped our approach to an educational platform combining AI literacy, C/C++ programming, and hands-on CubeSat activities. Rather than treating AI as a shortcut around learning technical skills, we are exploring how it can become another tool for developing them.
Using AI Is Not the Same as Understanding AI
Modern AI tools make sophisticated technology remarkably approachable. A student can ask a question in everyday language and receive an immediate, polished response.
That accessibility is one of AI's greatest strengths.
It can also create a problem.
An answer that sounds confident is not necessarily an answer that is correct. If students learn only how to ask AI for solutions, they may become better at obtaining output without becoming better at determining whether that output makes sense.
Our project starts from a different definition of AI literacy.
Students should learn how to interact meaningfully with AI systems, but they should also develop the critical-thinking skills necessary to evaluate what those systems produce.
Prompting is part of that process, not the end of it.
Prompting Is Really an Exercise in Defining a Problem
"Prompt engineering" can sound like a specialized AI discipline, but many of its most useful principles are much more fundamental.
A useful prompt requires someone to decide what they actually want.
What is the objective? What information matters? What constraints apply? What form should the result take?
A vague request leaves many of those decisions to the AI system. A precise request forces the person writing it to think through the problem first.
That makes prompting potentially useful in education for reasons that have little to do with memorizing special AI commands.
Students can learn that changing the quality of the input changes the quality and usefulness of the output. They can compare ambiguous instructions with specific ones. They can refine a request when the first response doesn't address the actual problem.
Those are AI skills, but they are also communication and problem-solving skills.
The Other Half of AI Literacy Is Evaluation
Getting a response should begin the next step of the process rather than end it.
Students need to ask:
Does this answer actually address the question?
Does the reasoning make sense?
Can I verify it?
What assumptions is the system making?
Do I understand the answer well enough to recognize when something looks wrong?
That last question is especially important.
If a student has no underlying understanding of a subject, it becomes much harder to evaluate an AI-generated response about that subject. The ability to use AI effectively is therefore connected to foundational knowledge rather than being a replacement for it.
This is one reason our educational concept combines AI literacy with programming.
Why We're Pairing AI With C/C++
The CubeSat project includes C/C++ programming as one of its core educational components.
The intention isn't to teach programming simply because programming is useful. It provides a counterbalance to the abstraction of modern AI tools.
AI allows a student to interact with an extremely complex system through natural language. Programming asks the student to be explicit.
A program has logic. Inputs produce results. Sensors return values that have to be handled. Conditions have to be defined. Data may need to be stored, displayed, transmitted, or compared.
If the code doesn't behave as expected, the student has a concrete problem to investigate.
That creates an opportunity for AI to support the learning process without becoming the learning process itself.
A student can use AI while working on code, but the code still has to run.
The sensor still has to return useful data.
The result still has to make sense.
Giving AI a Physical Context
We also wanted the programming environment to connect to something more tangible than exercises on a screen.
That led to the CubeSat component of the project.
The planned CubeSat hardware includes environmental sensing, acceleration sensing, air-quality sensing, camera input, local storage, an OLED display, Wi-Fi communication, and audio capabilities. Students can therefore work with software that interacts with physical inputs and outputs rather than only manipulating abstract values.
That distinction becomes particularly interesting when AI is introduced.
Suppose a program collects environmental information. AI can help a student think about how to structure the program or interpret the problem, but the underlying sensor reading exists independently of the AI's response.
If the software says one thing and the hardware does another, there is something to investigate.
That encourages a useful habit: test the output against evidence.
The same habit should apply when working with AI.
AI Should Support the Learning Loop
One of the planned components of the broader platform is an AI assistant alongside the lesson library and built-in code execution environment.
The design challenge is not simply adding an AI chat window.
The more important question is what role AI should play in the learning experience.
Our approach is centered on a loop:
Understand the problem → give clear instructions → evaluate the response → test the result → refine the approach.
That process applies naturally to both AI interaction and programming.
If an AI-generated suggestion leads to code that doesn't behave correctly, the student shouldn't simply ask for another complete answer. There is an opportunity to understand what happened, improve the question, examine the code, and try again.
The objective is not to prevent students from using AI to help solve problems.
It is to teach them what responsible problem solving with AI looks like.
Moving Beyond "AI Gave Me the Answer"
AI creates an unusual challenge for technical education because it can produce solutions to problems before a learner has developed the knowledge required to produce those solutions independently.
Simply banning the technology doesn't teach students how to use it.
Allowing unrestricted dependence on it doesn't solve the problem either.
A more useful educational goal is to make the student responsible for the final result.
AI can suggest.
AI can explain.
AI can help a student explore an unfamiliar concept.
But the student still needs to determine whether the result works and whether it makes sense.
Programming and physical computing provide a particularly useful environment for developing that mindset because they create feedback outside the AI system itself.
The program runs or it doesn't. A sensor produces data. A display shows a result. A physical device behaves in a way the student can observe.
There is something against which the AI's output can be tested.
What We're Learning
The central idea behind our AI-literacy work is simple: being comfortable with AI is not enough.
Students need the confidence to use these systems, but they also need enough technical understanding and critical thinking to challenge them.
That changes what AI education can look like.
Prompting becomes an exercise in defining problems clearly.
Evaluating responses becomes an exercise in critical thinking.
Programming provides a foundation for understanding logic and computational systems.
Physical hardware provides evidence that can be measured and tested.
AI then becomes part of a larger problem-solving process rather than an oracle sitting at the center of it.
For organizations thinking about AI education, that distinction matters. The long-term skill isn't knowing exactly how to interact with one generation of AI tools. Those interfaces will change.
The more durable skills are knowing how to define a problem, communicate requirements, evaluate information, test assumptions, and recognize when a result deserves another look.
Next Steps
Our CubeSat education project is intended to bring these pieces together through AI and coding lessons ranging from beginner to advanced, a built-in code execution environment, an AI assistant, and physical CubeSat hardware students can program and observe.
There is still development ahead across the hardware, software platform, lesson guides, and overall educational experience.
But the principle guiding that work is already clear.
We don't want to teach students simply how to get an answer from AI.
We want to help them develop the skills to decide whether it's a good one.