AI Coding Assistants for Test Authoring — quiz

A developer uses GitHub Copilot to generate Jest tests for a `calculateDiscount(price, code)` function. Copilot produces five tests, all asserting expected outputs for typical inputs. What is the most important review step before merging?

• Accept the tests as-is — Copilot output is always correct
• Verify the tests cover boundary conditions, negative paths, and invalid inputs (not just happy paths) and that assertions reflect intended behaviour, not just what the function currently returns
• Delete the tests because AI-generated tests are inherently invalid
• Run the tests once and merge if they pass