The EU AI Act entered into force in 2024. But knowing that changes nothing — acting on it does.
Ursprungligen var de fullständiga kraven på högriskssystem inom HR och recruitment planerade att träda i kraft den 2 augusti 2026. Efter en politisk överenskommelse i maj 2026 har deadlinen skjutits fram till den 2 december 2027.
That might sound like plenty of time. It is not.
Building AI readiness in an organization takes time — and the companies that start now will not be scrambling in 18 months.
To be honest: there is still a lot of confusion about what the regulation actually means in practice. What I can do is share what I know — and point to the questions every HR function should be asking right now.
What counts as high risk?
The AI Act divides AI systems into risk categories. What is directly relevant for HR is that systems used in employment decisions are classified as high risk.
That includes AI tools used to screen candidates, rank applicants, analyze interviews, assess performance, or otherwise influence selection decisions.
The requirements for high-risk systems include: transparency towards candidates, mandatory human review of AI-generated decisions, risk assessments, documentation of how systems work, and ongoing monitoring for bias and discrimination.
What that means in practice
If you use an ATS with automated sorting, you need to know how it makes its decisions — and be able to explain that to candidates.
If you use AI tools for personality assessment or interview analysis, you need to ensure that a human reviews and takes responsibility for the final decision.
Om ni inte vet vilka AI-system ni använder i er rekryteringsprocess – och det är vanligare än man tror – är det dags att ta reda på det nu. Inte i november 2027.
My take on AI in recruitment
I am not against AI in recruitment. Used well, it can free up time and bring better structure to the process.
But there are things AI cannot do. It cannot read whether the chemistry is right. It cannot judge whether this person will fit into your culture, with your managers, at this particular stage of your company's development.
Judgment cannot be automated. And it is still the most important tool in any recruitment process.
The EU AI Act reminds us of that. Frankly, it is a reasonable requirement.
What you should do now
• Map which AI tools are being used in your recruitment process
• Check whether your suppliers have documented compliance with the AI Act
• Ensure there is human review at every decision point
• Inform candidates about how AI is used in your process
• Eliminate any tools that analyse emotions or draw conclusions about personality traits — that is already prohibited
This is not a checklist for legal protection. It is a checklist for recruiting ethically and responsibly.
If you have thoughts, questions, or simply want to talk something through — feel free to get in touch. I am happy to have an initial conversation with no agenda.
Magdalena Hagström Ståhl
By M Consulting AB
Right person. Right place. Everything changes.
Sources:
European Commission — AI Act: official overview and implementation timeline
digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
DLA Piper – The Digital AI Omnibus: Proposed deferral of high-risk AI obligations under the AI Act (april 2026) knowledge.dlapiper.com
Crowell & Moring — Artificial Intelligence and Human Resources in the EU: a 2026 Legal Overview (February 2026)
crowell.com/en/insights/client-alerts/artificial-intelligence-and-human-resources-in-the-eu-a-2026-legal-overview
HireTruffle — EU AI Act and Hiring: 2026 Compliance Guide
hiretruffle.com/blog/eu-ai-act-hiring