Practise the language developers use when talking about what AI coding tools get wrong — incorrect suggestions, hallucinated APIs, and the importance of verifying AI output before shipping.
0 / 5 completed
1 / 5
A developer says: 'The Copilot ___ is incorrect — it's calling a method that doesn't exist on this object.'
'The suggestion is incorrect' is the direct way to say the AI's proposed code is wrong. It may look plausible — syntactically valid and logically structured — but produce incorrect results or reference things that don't exist.
2 / 5
The AI generated a call to `stripe.charges.createInstant()`. After checking the Stripe docs, you discover this is a ___ API — it was never real.
A hallucinated API is one the AI fabricated — often it sounds plausible and follows the naming conventions of the real library, making it easy to miss without checking the documentation. Always verify unfamiliar method calls against official sources.
3 / 5
Your tech lead's rule is: 'Always ___ AI output before you push.' What does this mean in practice?
Verifying AI output means actively checking that the code does what it claims — reading it critically, consulting official docs for any unfamiliar methods, running tests, and understanding the logic before shipping it.
4 / 5
Copilot suggests a React hook that was valid in React 16 but is deprecated in React 18. This is an example of ___ ___ ___ causing an incorrect suggestion.
AI models have a training data cutoff — they don't know about library updates, deprecations, or new APIs released after that date. For fast-moving ecosystems like React or cloud SDKs, always cross-check suggestions against the current documentation.
5 / 5
The concept of '___ ___ of AI tools' means a developer must review and own all AI-generated code they commit.
Responsible use of AI tools means understanding that AI-generated code is a starting point, not a finished product. The developer who commits the code owns it — and is accountable for its correctness, security, and maintainability.