Was your job application filtered by AI? How to request human review

Identify where automation affected the process

Automated recruitment is broader than a chatbot or recorded interview. A system may parse a CV, match keywords, rank applicants, score an online assessment or recommend who moves to the next stage. A person may still make the final decision, or the system may reject an application without meaningful human involvement.

Read the vacancy notice, application privacy notice, assessment invitation and rejection message. Look for the provider's name, the information analysed, whether profiling or automated decision-making is mentioned, and how the result is used. Save those notices because the wording can change after the campaign closes.

Check your own submission before blaming the model

Confirm that the uploaded file opened correctly, the application was complete and the stated location, work-authorisation and required qualifications matched your answers. A rejection after an eligibility question is different from a score based on a personality test or CV-ranking model.

Keep the version of the CV, answers, assessment confirmation, deadline and any technical error. Do not attempt to manipulate a system with invisible text or false keywords. Accurate, clearly structured evidence is easier for both software and human reviewers to assess.

Ask focused questions and request a real review

A useful request identifies the role and decision, asks whether automated processing materially influenced the outcome, and requests the available explanation, correction or human-review route. Point out a concrete issue such as an incorrect location answer, inaccessible assessment, missing qualification, file-processing error or fact the system appears to have misunderstood.

In March 2026, the UK Information Commissioner's Office told jobseekers that, when a decision is made without a real person involved, they should be told about the automation and can contest the decision, explain their position and request human review. The precise right and procedure depend on UK data-protection law and the facts; candidates elsewhere should consult their local regulator or privacy notice.

Do not overstate what the EU AI Act already guarantees

The European Commission lists AI used to analyse and filter applications or evaluate candidates as a potentially high-risk employment use case. However, following the 2026 AI Omnibus timeline, the high-risk rules for employment systems are scheduled to apply from 2 December 2027. A candidate should not assume that every planned high-risk obligation is already enforceable today.

Other rules may already apply, including data-protection, consumer, disability and anti-discrimination law. In the United States, the EEOC explains that tests and selection procedures can breach federal anti-discrimination law when they intentionally discriminate or disproportionately exclude a protected group without the required justification. Whether a particular tool or outcome is unlawful requires facts and jurisdiction-specific advice.

Use an automated-decision response checklist

• Save the vacancy, privacy notice, submitted CV, answers, assessment receipt and rejection message.
• Identify whether the issue is eligibility, a technical failure, a score or an automated ranking decision.
• Ask what tool or process materially influenced the result and whether a person reviewed it.
• Correct inaccurate personal data through the employer or provider's stated route.
• Request human review with one or two specific facts the automated process may have missed.
• Record the response and escalate through the privacy, accessibility or regulator route that applies locally.
• Continue other applications while a review is pending; a challenge does not guarantee reversal.

Review AI interview privacy and recording questions →Request an accessible assessment process →Browse current checked roles →