The EU AI Act is raising the bar for AI in recruitment. Significantly. But what does it mean for employers investing in AI-powered hiring tools? We sat down with Tengai's CTO Sebastian Otarola to break down the new legislation, what it means in practice, and the key things every company should know before investing in AI hiring software.
While AI can improve hiring efficiency and create a better candidate experience, recruitment is now officially classified as a high-risk area under the EU AI Act. That means employers using AI-powered recruitment providers carry a legal responsibility to ensure compliance, transparency and fairness throughout the hiring process. Getting it wrong is costly: non-compliance with the high-risk requirements can carry fines of up to €15 million or 3% of global annual turnover.
The good news? Choosing the right vendor is mostly about asking the right questions. Use the six below as your checklist the next time you evaluate AI screening software.
Sebastian Otarola, CTO at Tengai: AI systems that automatically rank, assess or make decisions about candidates without human involvement may fall under the EU AI Act's high-risk classification. This includes software using generative AI to evaluate candidates, screen CVs or automate hiring decisions.
Solutions built on validated psychometric frameworks, predefined scoring models and objective criteria — where AI is only used as decision support — are generally considered lower risk. In these cases, recruiters remain responsible for the final evaluation, ensuring true human oversight. And they must be able to show in detail how each decision and score was calculated.
Systems designed with transparent, traceable and controllable logic align much better with the requirements of the EU AI Act and are safer to use in hiring.
✅ Tengai's approach: Tengai uses validated psychometric frameworks and static scoring models with human oversight throughout the process. Generative AI is never used to make hiring decisions or rank candidates.
AI should never decide which job seekers move forward in a recruitment process. Recruiters or hiring managers must always be able to understand, question and override AI-generated results.
Make sure that:
AI systems must not discriminate based on gender, age, language, ethnicity or other irrelevant factors. RecTech vendors should be able to demonstrate how they continuously monitor and reduce bias.
Ask vendors:
How do you test fairness and data quality?
Are bias tests conducted regularly? How?
Do you use evidence-based frameworks or machine learning models?
What type of candidate data does the system access?
✅ Tengai's approach: Tengai uses validated assessments and static scoring frameworks designed to eliminate gender, age and ethnicity bias. The platform also avoids using historical hiring data. Only email addresses and interview voice recordings are stored after interviews.
Candidates have the right to know when and how AI is being used. Transparency is not only a legal requirement — it is critical for trust and candidate experience.
Ensure that:
Candidates are informed upfront about how AI is used
You can clearly explain which parts of the process are AI-supported
✅ Tengai's approach: Candidates are informed from the start that Tengai's AI avatar is used for interface and user experience purposes, not to evaluate performance, skills or competence.
Every score, how it was calculated and every decision must be documented and traceable. This is essential for legal compliance, explainability and accountability under the EU AI Act.
Check that:
All decisions are automatically logged
Raw data is stored for audits and reviews
Logs can be exported if requested by authorities
✅ Tengai's approach: Tengai stores all scoring data in raw format, making it fully traceable and available for audits or candidate reviews when necessary — with full explainability and all data stored within Europe.
High-risk AI systems must be secure, robust and resistant to manipulation. Employers should understand how hiring platforms and vendors test vulnerabilities and where data is stored.
Ensure that:
Systems are tested under real-world conditions
Protection exists against attacks and malicious inputs
The vendor complies with relevant ISO standards
Data is stored within the EU
✅ Tengai's approach: Tengai is certified according to ISO/IEC 27001, 27017, 27018 and 14001 standards. The platform is continuously tested for security and integrity, with all data stored in Sweden.
According to Sebastian, employers should be cautious if a vendor cannot clearly answer the above questions, or if the AI system:
Automatically ranks candidates using AI
Generate scores without human review
Lacks documentation, transparency or bias control
These are strong indicators that the AI hiring solution may not comply with the EU AI Act.
A compliant recruitment system helps companies make faster and more objective hiring decisions without removing human oversight from the process.
As the EU AI Act becomes fully enforced, ethical AI will not just be about compliance. It will become a competitive advantage.
Want to see what compliant AI hiring looks like in practice? Read how the trade union Vision reduced time-to-hire while improving candidate experience with Tengai.
With Tengai, companies get a screening and hiring platform designed to meet the requirements of the EU AI Act from day one.
Book a demo to see how Tengai meets every requirement.