When meeting with prospective AI recruitment vendors use these questions to ensure they are not only compliant but their use of AI is also ethically fair and responsible. Consider adding these questions to your RFP (request for proposal).

This checklist will help you:

– Understand not just what a vendor’s AI tool does, but how it works.

– Distinguish between superficial automation and science-backed, compliant solutions.

– Identify red flags such as opaque, inconsistent, or LLM-driven scoring systems.

How is your AI system classified under the EU AI Act? (e.g. high-risk, minimal risk, prohibited) 

Have you completed a formal risk assessment for your system in the EU context? 

Can you share documentation outlining your classification rationale and impact assessment?

What information is made available to end users and candidates about how your AI works? 

Do you provide clear documentation on AI decision logic and outputs? 

Can users override or contest AI-driven decisions? 

Do you allow human oversight of automated decisions?

‍

How do you fulfill your obligations as a Provider under the EU AI Act? 

Do you provide documentation and support to help Deployers (e.g. recruitment agencies or employers) meet their obligations? 

Do you have a Data Protection Officer (DPO) or an appointed AI Officer responsible for governance?

‍

What steps do you take to ensure your AI does not introduce or amplify bias?

Have you audited your models for demographic fairness and explainability? 

Do you conduct regular fairness and accuracy evaluations across diverse candidate pools?

‍

Does your system use facial recognition, biometric data, or emotion recognition? (If so, beware: these may fall into prohibited or high-risk categories) 

Do you engage in profiling, scoring, or classification that could affect candidates’ rights? 

Have you conducted a Fundamental Rights Impact Assessment (FRIA)?

‍

How is candidate data collected, stored, and used? 

Do you provide mechanisms to obtain and manage informed consent? 

Can candidates request deletion or correction of their data?

Can you provide compliance documents or certifications (e.g. CE mark, conformity assessment, etc.)? 

Do you maintain logs and traceability of model outputs and updates? 

Do you offer support to customers preparing for audits or regulator reviews?

‍

Want to learn more about ethical AI in recruitment and what compliance with the EU AI Act looks like in practice?

Watch our on-demand webinar with Alex Tidgård (CPO, Jobylon) and Fredrik Törn (CPO, Hubert) here.

Together with Jobylon, we’ve created an extended checklist of questions to ask AI recruitment software vendors. It covers key areas like Candidate Assessment and Scoring, Data Rights, Bias Mitigation, Transparency, and Training Data Practices to help you evaluate tools responsibly.

‍

Download the full checklist for free here.