Rubber-stamping
Rubber-stamping is approving a decision without really reviewing it. In AI hiring it means a recruiter clicks “approve” on the AI’s recommendations without checking each case. European courts have treated that kind of sign-off as an automated decision, not as human oversight.
Rubber-stamping rarely starts on purpose. A tool sorts 200 candidates, the recruiter has an hour and the "approve all" button is right there. The recruiter is in the loop on paper. In practice the AI decided.
Why is a human click not enough?
Many teams assume a decision is not automated as long as a person signs off at the end. European courts have said otherwise.
- SCHUFA (Court of Justice of the EU, C-634/21, 7 December 2023). When a person gives an algorithmic score a "determining role", the decision counts as automated under GDPR Article 22, whoever clicks the button.
- Amsterdam Court of Appeal (4 April 2023). Dismissals of Uber drivers were approved by people in name only. The court found that purely symbolic human validation breaks GDPR Article 22.
Applied to hiring, a recruiter who approves hundreds of AI-recommended candidates a day, without looking at each case, is rubber-stamping. The EU AI Act adds its own demand for human oversight on top.
How does rubber-stamping differ from real human oversight?
| Rubber-stamping | Real human oversight | |
|---|---|---|
| Time per candidate | Seconds or a bulk click | Enough to read the answers |
| What the reviewer sees | A score or a yes or no | The answers and the reasoning behind the result |
| Authority | Can only confirm | Can change the outcome |
| Record | None | A logged reason when the reviewer overrides |
| Who decides | The AI, in practice | The person |
How do you spot rubber-stamping in your own process?
Look for these signs:
- Recruiters approve almost every AI recommendation.
- A review takes a few seconds per candidate.
- The reviewer sees only a score, not the answers.
- Nobody can remember the last override.
- There is no log of who changed what and why.
Several of these signs together usually point to rubber-stamping.
What does real oversight look like?
- Time per case. Plan the reviewer's day so each candidate gets a real look.
- The evidence in view. Show the answers and the reasoning, not only a number. See candidate scoring.
- Authority to override. The reviewer can change the outcome without asking permission.
- A record. Log each override with a reason and keep the logs for at least six months.
- No automatic rejection. Let a person decide every no, especially on knock-out questions.
How does Ringtime support real oversight?
Ringtime's AI agents run the screening conversation. A person keeps the decision.
- The assistant never rejects a candidate by itself.
- A candidate who fails a must-have question goes to recruiter review.
- Every result comes with the candidate's answers, so the recruiter sees why.
- Recruiters can read the full conversation in the Ringtime web app.
- The result also lands in your applicant tracking system (ATS), where the recruiter takes the next step.
This page is general information, not legal advice. Talk to your legal team or data protection officer about your own setup.
To see where the human click stops counting, read what the EU AI Act means for staffing.
