Using AI in your job search without losing your voice

Recruiters now read hundreds of AI-assisted applications a week and have become good at spotting the generic ones. That does not mean avoid the tools — it means use them where they have real leverage and keep a human hand on the facts.

Where AI helps most

• Extracting requirements and repeated keywords from a long job description.
• Rewriting weak bullets into action-scope-outcome form using your own inputs.
• Producing a first draft of a cover letter you then cut by a third.
• Generating interview questions and pressure-testing your answers out loud.
• Proofreading for consistency in tense, dates and formatting.

Where it hurts

• Inventing metrics, tools or responsibilities you never had.
• Flattening your voice into corporate filler — 'leveraged synergies', 'passionate about excellence'.
• Mass-applying with unread output, which burns companies you may want later.
• Assistance during a live interview, which is increasingly detected and is usually an instant disqualification.

Embellish the description, never the record. Every line has to survive a reference check and a follow-up question.

A safe workflow

1. Write your raw history yourself, ugly and factual.
2. Feed the posting and your raw history in together, and ask for rewriting only.
3. Read every generated line and delete anything you cannot personally defend.
4. Restore two or three phrases that sound like you actually talk.
5. Check names, dates, titles and numbers against the original by hand.

Disclosure

Nobody expects you to declare that a tool helped you format bullets, just as nobody declares spellcheck. If an employer explicitly asks for unassisted work on a take-home task, honour that exactly — a caught violation ends the process.