When should you NOT use ML/AI for a fresher project?

When skip ML

For When should you NOT use ML/AI for a fresher project, interviewers listen for a crisp definition, how the pieces interact in a real system, and one trade-off or failure mode you have seen or would guard against.

• Small labeled data and simple rules suffice.
• Need 100% explainability (regulations).
• Problem is CRUD/reporting not prediction.
• Cost/latency of ML ops not justified.