# Test an AI Coding Tutor with a Bug You Can Explain

A coding tutor can produce a correct answer while teaching very little. To compare learning tools, give each one a small bug and ask for a hint before allowing a complete solution.

## Prepare a controlled task

Use a short loop that skips an item or accesses an index outside a list. Write down the expected output yourself. Keep the language and prompt identical across candidates, so differences in the task do not dominate the comparison.

The tutor checklist at [EasyGuides](https://easyguides.com/technology/how-to-choose-an-ai-coding-tutor/) suggests examining explanations, practice and debugging behavior. Turn those categories into observations: did the tool ask what you expected, identify the relevant concept and leave room for another attempt?

## Check the explanation independently

For a Python exercise, use the [official Python tutorial](https://docs.python.org/3/tutorial/) to check the underlying language behavior. Run the corrected program and vary one input. A response that sounds confident still needs to match what the interpreter does.

## Measure what you can do afterward

Close the chat and solve a nearby problem without copying. Explain why the boundary condition failed in your own words. If you cannot, ask for a smaller exercise or a clearer explanation. Keep notes about that learning step rather than ranking tutors by how quickly they reveal the finished code.

Try an empty input and a one-item input as well. Ask the tutor to predict each result before running the code. Boundary cases reveal whether its explanation describes the program or merely repeats the corrected snippet.