Artificial intelligence has made its way as a tool in schools. Whether it’s asking for clarifications on topics to generating images for projects, it’s influence on education has shaped the way students learn and engage with new ideas in the classroom. As students engage with a variety of AI platforms, including ChatGPT, Gemini and Copilot, the rise of AI usage in the classroom is prominent. However, even with new technological advancements, artificial intelligence still tends to miss the mark. From algorithmic bias to disinformation, AI can negatively impact learning in and outside of the classroom.

The use of AI is evident and present in the classroom. But, as an educator, when and what could I use it for?

Here are some ChatGPT Do’s and Don’ts. Use these guidelines to rethink and reshape the use of AI in the classroom in computer science education.

Use AI to:

• Debug after trying first: Use AI to help identify errors, logic errors, or runtime errors after students have tested their code and attempted multiple fixes.

• Set firm guidelines: Clearly define how AI may be used during algorithm design, pseudocode writing, debugging, and testing—and when AI use is not permitted.

• Organize ideas: Use AI to help organize steps in an algorithm, outline pseudocode, or break a problem into smaller parts.

• Generate practice code: AI can generate code examples or partially completed programs for students to trace, test, and debug.

• Provide feedback: AI can support feedback on program logic, use of variables, conditionals, loops, and clarity of comments.

• Practice tracing code: Ask AI to create code where students track variable values, follow loops and conditionals, and predict outputs before running the program.

• Learn coding vocabulary: AI can explain CS vocabulary such as variables, conditionals, loops, algorithms, data types, and events in student-friendly language.

• Test programs: Use AI to suggest test cases or edge cases to help students check whether their programs work as intended.

• Reflect on learning: Ask AI questions like, “Why does this algorithm work?”, “What part of the program controls the flow?”, or “How could this solution be made more efficient or readable?”

• Support accessibility: AI can help rephrase instructions, explain errors, or provide alternative explanations for CS concepts.

Do NOT use AI to:

• Find information without verification: AI may generate incorrect explanations of algorithms, misleading code, or examples that do not work. All code and concepts must be tested and verified.

• Answer coding challenges: AI should not generate solutions to programming tasks. Give students the chance to test code.

• Write the entire program: AI should not be used to produce full programs, final project code, or completed algorithms.

• Replace algorithm design: Students should not use AI to create algorithms, pseudocode, or flowcharts that they did not design themselves.

• Skip debugging and testing: Debugging, testing, and revising programs are essential parts of the CS learning process.

• Assume AI is always correct: AI-generated code may contain logic errors, inefficient solutions, or poor programming practices.

• Avoid learning core CS concepts: Students still need to learn how variables, loops, conditionals, events, and data work without relying on AI.

• Use AI to bypass collaboration: AI should not replace peer discussion, pair programming, or explaining ideas to others.

It’s clear: Artificial intelligence can be a valuable support tool in middle school computer science when used intentionally and with clear expectations. Using these guidelines AI helps can be used to help students learn. It can debug, organize ideas, and reflect on learning without replacing critical thinking or problem-solving. When used responsibly, AI can enhance learning while reinforcing essential computer science skills.