How we stay right
Sources
Every claim on this site that rests on research or law is listed here, with what the source found and how far it goes. One study is one study. We say "can be", not "is".
"Meritocracy doesn't happen by declaring it. It happens by measuring it."
Castilla, E. J., & Benard, S. (2010). The Paradox of Meritocracy in Organizations. Administrative Science Quarterly, 55(4), 543–576. Publisher · MIT open access
What it found. Three experiments with 445 participants who had managerial experience. When the company's stated values emphasised meritocracy, managers gave larger rewards to men than to equally performing women.
How far it goes. Experiments on reward decisions, not a field study of hiring. It supports the point that declaring merit is not enough. It does not measure any particular company.
"Hiring AI can be fairer than people, but only when someone checks."
Cowgill, B. Bias and Productivity in Humans and Algorithms: Theory and Evidence from Résumé Screening. Working paper.
What it found. A field experiment at one employer. The candidates picked by the algorithm but not by human screeners were more likely to pass interviews and receive offers, and the algorithm selected more non-traditional candidates.
How far it goes. One experiment, one employer, a working paper. It shows that an algorithm can do better than human screening when its results are measured. It does not show that hiring AI is fairer in general, and we never say so.
"AI shouldn't be at the helm of HR. People should, with evidence."
Regulation (EU) 2024/1689 (the AI Act), Article 14, human oversight of high-risk AI systems, and Article 26(2), which requires those who deploy them to assign oversight to people with the competence and authority to exercise it. Official text
How far it goes. This is a legal requirement, not an opinion. Hiring and worker-management systems are high-risk under Annex III. The Digital Omnibus deferred the high-risk obligations, Article 14 included, to 2 December 2027. What binds today for hiring AI is the Article 5(1)(f) prohibition on emotion inference at work, in force since 2 February 2025, and the Article 50 transparency duties.
How we use sources
- A claim appears on this site only if its source is listed here.
- Numbers are quoted as the paper reports them, with the year.
- Studies from one country or one employer are described as such.
- Found a claim we have stretched? Write to stela@fairaudit.ai. We would rather fix it than defend it.
AI shouldn’t be at the helm of HR. People should, with evidence.
The EU AI Act requires human oversight of high-risk hiring systems (Art. 14). We help you show it is real.
The assistant: sign in with your email. We confirm access by email. The high-risk obligations for hiring systems, Article 14 included, apply from 2 December 2027. Regulatory information, not legal advice.