Is AI candidate scoring "high-risk"?

Yes. The AI Act (EU Regulation 2024/1689) classifies candidate screening in Annex III (high-risk). aiKip applies transparency, human oversight and documented criteria.

Can aiKip reject a candidate automatically?

No. The AI proposes a score and a summary; the final decision always stays with the human recruiter.

Artificial intelligence transparency

aiKip uses artificial intelligence to assist recruiters in analyzing applications. This document describes what the AI does, what it does not do, and candidates' rights — in line with the European AI regulation (AI Act, EU Regulation 2024/1689), which covers AI systems used in employment (Annex III).

What the AI does

CV extraction: information from the CV (skills, experience, education) is extracted automatically to pre-fill the candidate profile.
Match score: each application receives an indicative score from 0 to 100 comparing the profile to the job, against four weighted criteria: technical skills (35%), professional experience (35%), education (15%), motivation (15%).
Factual summary: a 2-3 sentence summary, strengths and watch points are generated to speed up reading the file.

What the AI does not do

No automated decision: the AI does not reject, shortlist or definitively rank any application. Every stage change (shortlisting, interview, rejection, hiring) is a manual action by an identified recruiter.
No prohibited criteria: the system neither receives nor evaluates age, gender, origin, precise address, photo or any other discriminatory criterion.
No data sent outside the EU: the model used is a French model, hosted in France. See our sovereignty commitment.

Human oversight

The score is indicative and presented as such in the recruiter interface. Recruiters can recompute a score, re-analyze a CV, and make decisions alone. Companies using aiKip undertake (Terms of Use) not to base a hiring decision on the score alone.

Traceability

Each scoring and each stage change is logged (timestamp, author). The full history of an application can be viewed by the recruiter and is kept for the lifetime of the file (24 months maximum, then automatic anonymization).

Your rights as a candidate

Information: the application form mentions the AI-assisted analysis before any submission.
Human intervention: you can request that a recruiter review your file without taking the score into account (art. 22 GDPR), via the messaging in your candidate space or our contact form.
Contestation: you can contest a decision and request explanations about the use of AI in processing your application.
Erasure: you can withdraw your application at any time; your data is then processed according to our GDPR policy.

Roles and AI Act compliance

Under the AI Act, aiKip is the provider of the AI system and each recruiter using it is a deployer. As the system falls under Annex III (high-risk, employment), aiKip maintains the corresponding measures — designed-in human oversight, risk management, technical documentation, automatic logging, and transparency to deployers and candidates — and is completing the conformity work required by the regulation, whose high-risk obligations apply from 2 August 2026. Prohibited practices (e.g. emotion recognition in the workplace) are not implemented.

Known limits

An AI score can be wrong: an atypical CV, a career change or an unusual document format may be poorly evaluated. This is precisely why the score is never a decision — only a starting point for the human reading of the file.