Inventory
Map AI features across ATS, sourcing and interview tools.
AI compliance for recruitment and staffingWHEN AI AFFECTS
WHEN AI AFFECTS
WHO GETS SEEN.
Blanche helps recruitment agencies and HR teams identify AI inside their technology, distinguish drafting support from candidate decision systems and put meaningful governance around higher-impact use.
01ATS and vendor review02Candidate-impact assessment03Meaningful human oversight04Recruiter AI rules
When can recruitment AI be high-risk under the EU AI Act?
Certain AI systems used to recruit or select people, filter applications, evaluate candidates, make work-related decisions or monitor workers can be high-risk when the legal definition and relevant criteria are met.
01 / Find the AI
The important features may be inside existing software.
A recruitment business may not develop AI but can still use it through an applicant tracking system, sourcing platform, interview tool, ranking feature or productivity assistant. Vendor terminology does not always make the level of automation clear.
The inventory should identify which features analyse, filter, score, rank, recommend or monitor people, and how recruiters use those outputs in practice.
02 / Separate the use cases
Drafting a job advert is not ranking a candidate.
Generative AI used to improve wording, summarise internal notes or draft routine communication does not automatically become high-risk. The compliance position changes when an AI system influences access to employment, progression, allocation of work, performance evaluation or termination.
Those uses need closer AI Act classification, GDPR, equality, transparency and human oversight attention.
- Automated CV filtering or ranking
- Candidate scoring and recommendation
- Video or interview analysis
- Targeted job advertising
- Employee monitoring or performance evaluation
- Shift, task, promotion or termination recommendations
03 / Keep a person responsible
Human review must be able to change the outcome.
A recruiter should understand the system's purpose and limits, have enough information to question the recommendation and hold genuine authority to disregard or reverse it. Routine approval of a score is not meaningful oversight.
Candidates also need an appropriate route for information, correction and challenge where decisions or personal data processing affect them.
Clear sequence. Clear ownership.
Classify
Screen intended purposes and candidate or worker impact.
Control
Set human review, data, transparency and vendor rules.
Train
Give recruiters role-specific examples and escalation routes.
Common questions
Useful answers before you start.
01Is every ATS an AI system?+
No. An ATS may include conventional automation, AI features or both. Each relevant feature and intended purpose should be assessed against the AI Act definition.
02Does human approval make candidate ranking low-risk?+
Not automatically. The system may still be high-risk based on its intended purpose, and the human review must be competent, informed and able to change the outcome.
03Can recruiters use ChatGPT for job descriptions?+
It may be possible with approved tools and suitable rules. Staff should protect personal and confidential data, check discriminatory language and verify the final content.
04What should we ask an AI recruitment vendor?+
Ask which features use AI, their intended purpose, training and evaluation information, known limitations, data flows, logging, human oversight support, change notices and AI Act role documentation.
Need practical rules for staff AI use?
Ready when you are
Ready to put this in place?
Tell us briefly how your organisation uses AI. We will review the context and confirm the most useful next step before discussing scope.
No assessment or estimate is required before you enquire.