How to Debug Code With AI

Quick Answer

Paste the broken code, the exact error message, and a description of what it should do. Ask AI to identify the bug, explain why it happens, and provide a fixed version. Most straightforward bugs take under 5 minutes to resolve.

What This Workflow Helps You Do

• Fix error-message bugs faster
• Identify logic errors that don’t throw errors but produce wrong results
• Understand why a bug happened, not just how to fix it
• Get unstuck when you’ve been staring at the same code too long

Step-by-Step Process

Step 1: Fix an error message bug (5 minutes)

The most straightforward case. You have an error and a stack trace.

I have a bug in my [language] code. Here is the error:

Error message:
[paste the full error message and stack trace]

Relevant code:
[paste the function or block where the error occurs]

What I expected to happen: [describe expected behavior]
What actually happens: [describe actual behavior]

Please:
1. Identify the cause of the error
2. Explain why it's happening
3. Provide a corrected version of the code
4. Note anything else in the code that looks like it could cause problems

Step 2: Debug a logic bug (no error message)

Harder, the code runs but produces wrong results.

My [language] code runs without errors but produces incorrect output.

Code:
[paste the relevant function or block]

Input I'm testing with: [describe or paste test input]
Expected output: [what it should produce]
Actual output: [what it actually produces]

Walk through the logic step by step to find where the output goes wrong.
Then provide a corrected version.

Step 3: Get a full code review for hidden bugs

Review this [language] code for bugs, edge cases, and potential issues.

[paste the code]

Check for:
1. Logic errors or off-by-one mistakes
2. Edge cases that aren't handled (null values, empty arrays, division by zero)
3. Any security issues (if applicable)
4. Performance problems for large inputs
5. Anything that would behave unexpectedly in production

For each issue found: explain the problem, the risk, and the fix.

Step 4: Understand a bug you’ve already fixed

I had this bug in my code: [describe the bug]
I fixed it by: [describe the fix]

Can you explain:
1. Why the original code caused this bug
2. Why my fix works
3. Whether my fix could cause any other issues
4. If there's a cleaner or more robust way to solve it

Step 5: Rubber duck debugging with AI

When you’re stuck and don’t even know where to start:

I'm stuck debugging this [language] code. I'll describe the problem and I want you to ask me clarifying questions to help me find the bug.

Context: [describe what the code is supposed to do]
The problem: [describe the symptoms]
What I've already tried: [list what you've ruled out]

Ask me questions one at a time to narrow down where the bug might be.

Debugging Checklist

• Paste the exact error message (not a paraphrase)
• Include the relevant code (not the entire codebase)
• Describe expected vs. actual behavior
• Test the suggested fix before accepting it
• Understand why the fix works before moving on
• Check for related issues in similar code

Common Mistakes

Pasting too much code. Give AI the relevant function or block, not your entire file. Focused context produces better debugging.

Paraphrasing the error. Always paste the full, exact error message. Details in the message often point directly to the cause.

Accepting the fix without understanding it. If you don’t understand why the fix works, you’ll hit the same bug again. Ask AI to explain the fix.

Not testing edge cases. Ask AI to identify edge cases for the fixed code, not just that it passes your current test.