AI Programming Policy

We define AI Slop Programming as code produced by AI that the author cannot explain, test, or maintain. It is often characterized by inconsistent style, wrong assumptions, repetitive code, and “looks right” changes.

The field of Artificial Intelligence has made incredible strides in recent years. With that has come challenges in defining acceptable use at YETI. We cannot feasibly ban the use of AI during the software development process. As such, this policy seeks to clarify how YETI’s web and robot programming mentors view AI usage. We encourage all team members to experiment with this powerful tool and become familiar with its strengths and weaknesses.

Core principles:

• We want to shape programmers who can think critically, regardless of AI use.

• When used as a partner, AI can be a powerful enhancement to productivity. When used carelessly, it hinders learning and understanding.

• It is the responsibility of all programmers to understand all code they commit. If you don't understand the code you check in, when something goes wrong, you will not know why.

• We’d much rather see slower, thoughtful progress than reckless AI-driven development.

• Mentors’ time is valuable. Spending time reviewing primarily AI-generated PRs is not a good use of our time.

What’s allowed?

• Explaining concepts, errors, and snippets of code

• Generating small snippets of code that you understand line-by-line

• Generating documentation from code

What’s not allowed?

• Submitting large, primarily AI-written PRs

• Random additions of libraries, new patterns, etc., without a functional reason

• Using AI for assigned learning tasks (e.g., Codecademy, React-Tac-Toe/Web Challenge)

Before submitting a PR

• You understand what the code is doing and why

• Check, can you explain the PR clearly to a mentor in 2 minutes?

If you used AI meaningfully, please disclose its usage in a PR description. If we suspect a PR is AI-generated, we will close the PR and ask you redo the work without AI assistance.