Poland's Data Protection Chief Demands Urgent AI Rules for Hiring

Mirosław Wróblewski, president of Poland's Data Protection Authority (UODO, Urząd Ochrony Danych Osobowych), sent a formal letter on July 16, 2026 to the Ministry of Family, Labor and Social Policy demanding urgent regulation of artificial intelligence systems used in recruitment and hiring. The letter targets tools that mass-screen CVs, conduct preliminary interviews with candidates, and tell employers whom to hire.

What the UODO chief did

Wróblewski's letter is a formal request addressed to the ministry responsible for labor law, not an administrative decision or a penalty. UODO has no power to draft Labor Code provisions on its own, so it chose to appeal to the ministry in hopes its demands will be folded into ongoing legislative work. The authority has sent similar letters before on general AI policy in Poland, but this is the first one dealing directly with recruitment and hiring.

The intervention follows complaints reaching the authority from candidates and employees who question how algorithms evaluate their applications. UODO stresses that the use of AI in HR processes is growing faster than the data-protection rules meant to keep pace with it.

What risks the authority identifies

The UODO president points to several specific risk mechanisms. The first is the entrenchment of biases embedded in training data, which can lead AI systems to systematically rate certain groups of candidates lower without disclosing the logic behind their decisions. The second is large-scale processing of personal data, including special category data such as health, orientation, or belief information, which is subject to especially strict protection in recruitment.

The third risk is that algorithms may analyze more information about a candidate than Article 22(1) of the Labor Code allows, a provision that precisely limits what data an employer can request during recruitment. The fourth problem is the circumvention of GDPR safeguards against fully automated decision-making about a person, in cases where an AI system effectively determines the recruitment outcome while human involvement in the decision is merely nominal.

AI systems used in recruitment have been classified as so-called high-risk systems - statement from Poland's Data Protection Authority (UODO)

What UODO is demanding

The authority is calling for additional safeguards to be written into labor law to protect candidates and employees evaluated by AI. Among its concrete proposals: a mandatory data protection impact assessment before such systems are deployed, transparency standards for tools classified as high-risk, and clear rules defining the system's purpose, the data controller's role, and the rights of the person being evaluated.

UODO has offered the ministry expert support in assessing whether future regulations comply with GDPR, signaling a willingness to cooperate on legislation rather than merely raising objections. The approach is meant to improve the odds that its demands make it into an actual draft law rather than remaining just a declaration.

What it means for Polish companies and candidates

For HR departments using AI tools to screen CVs, the letter signals that current Labor Code provisions may soon be clarified specifically around automated candidate analysis. Companies already using such systems should be prepared to demonstrate what data is being processed, for what purpose, and whether the final decision genuinely rests with a human.

For candidates, this could mean greater transparency in the recruitment process and the ability to demand an explanation of why an algorithm rejected their application. UODO's letter fits into the broader regulatory push around the EU's AI Act, whose provisions on high-risk systems, including those used in recruitment, are being phased in through 2027.

The Ministry of Family, Labor and Social Policy's response is not yet known. The pace of further work will depend on whether UODO's demands are folded into one of the Labor Code amendment drafts already underway, or whether they will require a separate legislative initiative.