AI and recruitment: why your 40 applications all look the same

Employer
Mélissa TessierMélissa Tessier · September 1, 2026 · 5 min read

Try it tonight. Print the last ten applications you received for your open role. Hide the names. Try to guess who is who.

You won’t manage it. Everyone is “rigorous, motivated and a great team player.” Three letters open on the same sentence. Two different backgrounds tell exactly the same story, in the same words.

This isn’t a feeling. It’s the new face of AI and recruitment: in France, half of candidates now use artificial intelligence in their job search, and among those under 30 it’s nearly two in three. The number one use? The cover letter. Quebec is following the same curve.

AI didn’t create the problem, it finished it

Let’s be honest: applications already looked alike before ChatGPT.

The résumé always pushed toward uniformity. The same downloaded templates, the same formulas recommended by the same guides, the same keywords placed to clear the same filters. We asked different people to describe themselves in an identical format, then acted surprised not to see any differences.

Generative AI simply took that logic to its conclusion. It produces the perfectly optimized application: no typos, well structured, tailored to the posting. For everyone. Hence the paradox: every application is better, and the whole pile is worth nothing.

If a machine can write a perfect application in twelve seconds, the written application never measured much.

The reflex, when you suspect a generated letter, is to feel cheated. “They didn’t even take the time to write it themselves.”

Flip the perspective. This candidate faces postings that demand precise keywords, screening systems that eliminate in six seconds, and applications that, nine times out of ten, never get a reply. In that game, spending an hour crafting a handmade letter no one will read is the irrational choice. Optimizing with AI is playing smart by rules they didn’t write.

This is where we are: letters generated by machines, sorted by machines. The bots talk to each other while the candidate and the employer still haven’t met.

AI and recruitment: the race you can’t win

The false good idea is detection. In a Robert Half survey of more than 1,500 hiring managers, 79% say they can spot AI in an application. Detection tools get it wrong regularly, and they mostly penalize people who simply write well, or those for whom the language is a second one.

But suppose detection worked perfectly. What would you actually know? Who used ChatGPT. Not who can handle an upset client. Not who will get along with your team. Not who will still be there in a year.

Detecting isn’t sorting. Even won, this race leads nowhere.

Ask for what can’t be generated

The only lasting answer is to change the field.

Stop asking candidates for a document a machine writes better than they do. Ask them for something no one can do in their place: to introduce themselves. Live, with their voice, their energy, their way of explaining what they’ve done and what they’re looking for.

A letter is generated. A way of being is shown. Thirty seconds of video tell you more than forty perfect letters, precisely because video is the only format where the person can’t be replaced by their tool.

This is the field Joblinks has chosen: a local platform where candidates introduce themselves on video and where companies show their team and their culture. The first impression becomes human again, in both directions. AI killed text-based sorting? Good. It wasn’t sorting anything.

The question is no longer “how do I know if this letter was generated?”. It’s: what am I asking my candidates that ChatGPT can’t do in their place?

Your next candidates are ready to show up for real

Meet them on Joblinks: video profiles, a team and a culture you can see, a first impression that’s finally human.

Sources and references

The figures cited come from the sources below. We invite our readers to consult them to verify and explore further.

  • Hellowork: April to May 2025 survey (2,247 candidates, 486 HR professionals). 50% of candidates use generative AI in their job search; 72% for the cover letter; 63% among 16 to 29 year-olds. helloworkplace.fr
  • Robert Half Canada: survey of more than 1,500 hiring managers. 79% say they can detect AI use in an application. roberthalf.com
Was this article helpful?

Partager l'article

Share the article

Rejoins la communauté Joblinks

Join the Joblinks community

Que vous cherchiez un emploi ou la bonne personne pour votre équipe, recevez des conseils concrets, des outils pratiques et les nouveautés de la plateforme. Une infolettre utile, jamais envahissante.

Whether you’re looking for a job or the right person for your team, get concrete tips, practical tools and platform updates. A useful newsletter, never intrusive.

Suivez-nous sur LinkedIn

Follow us on LinkedIn

Soyez à l’affût de nos derniers développements et actualités.

Stay tuned for our latest updates and news.

Articles connexes

Related articles

Holiday hiring: your real deadline isn’t Christmas

Employer
Ophelie WelldonOphelie Welldon · 6 min read

What an incomplete profile says to the other side

CandidateEmployer
Ophelie WelldonOphelie Welldon · 6 min read
Relire une offre d’emploi ligne par ligne

Your Job Posting Tells the Story of Your Bad Hires

“Must be autonomous.” That line added in ten seconds says a lot — not about the role, but about your last bad hire.
Employer
Mélissa TessierMélissa Tessier · 7 min read
Afficher le salaire dans une offre d’emploi

Posting the Salary: The Honest Math

Posting the salary is scary for three good reasons. Put them up against what silence really costs you.
Employer
Mélissa TessierMélissa Tessier · 5 min read
250 job applications piling up in an inbox

250 Applications, No Information

You’ll never read your 250 applications. That’s no longer a flaw in your method — it’s the sign that a tool has stopped working.
Employer
Mélissa TessierMélissa Tessier · 5 min read