Short answer
AI used to recruit and select people is high-risk under the EU AI Act (Annex III). After the Digital Omnibus (Regulation (EU) 2026/1744), the high-risk requirements apply from 2 December 2027. Already in force are the bans on certain AI practices and the AI literacy requirement (since 2 February 2025) and the transparency rules in Article 50 (since 2 August 2026). As a deployer you must, among other things, ensure human oversight, follow the provider's instructions and inform the people affected. This is information, not legal advice.
On this page
The EU AI Act is the world's first broad law on AI, and recruitment is one of the areas it regulates most strictly. For recruiters, that means new requirements on the tools you use and on how you use them. Here we go through what applies, when it applies and what you can do already now.
Why recruitment counts as high-risk
The AI Act sorts AI systems by risk. Some practices are banned, some systems are high-risk and come with strict requirements, and the rest have lighter requirements or none at all.
Recruitment falls into the high-risk group. Under Annex III, AI systems intended to be used to recruit or select people are high-risk, for example systems that target job ads, analyse and filter applications or evaluate candidates. The same goes for AI used for decisions on promotion, termination or task allocation, or to monitor and evaluate employees' performance.
The reason is that such systems can affect people's chances of getting a job and earning a living, and that they can carry over biases in data, for example related to gender, age or origin.
The timeline: what applies when?
The regulation entered into force in 2024 but applies in stages. Three dates matter for recruitment:
| Date | What starts to apply | What it means for recruitment |
|---|---|---|
| 2 February 2025 | Prohibited AI practices and AI literacy (Articles 4 and 5) | Some practices may not be used at all. Staff using AI must have sufficient knowledge of it. |
| 2 August 2026 | Transparency obligations (Article 50) | People must be told when they interact with an AI system, and AI-generated content must be recognisable. |
| 2 December 2027 | Requirements for high-risk AI in Annex III | Full requirements on AI systems for recruitment, for both providers and deployers. |
The high-risk requirements were originally due to apply from August 2026. Through the Digital Omnibus, Regulation (EU) 2026/1744, they were moved to 2 December 2027 for Annex III. That gives more time, but no more leeway: the requirements are the same.
What already applies
Prohibited AI practices
Since 2 February 2025, certain AI practices are banned. Two are particularly relevant to recruitment:
- Emotion recognition in the workplace. AI that infers the emotions of people in the workplace, for example in video interviews, is banned except for medical or safety reasons.
- Biometric categorisation of sensitive traits. AI that uses biometric data to infer, for example, ethnic origin, political opinions, religion or sexual orientation is banned.
AI literacy
Since the same date, companies that use AI must make sure their staff have sufficient AI literacy. For a recruiting team, that in practice means that the people using AI tools understand what the tool does, what it doesn't do and where it can get things wrong.
Transparency
Since 2 August 2026, the transparency obligations in Article 50 apply. Anyone interacting with an AI system, for example a chatbot on a careers page, must be told so, and AI-generated content must be recognisable as such.
What you need to do as a deployer from December 2027
The regulation distinguishes between the provider, who develops the AI system, and the deployer, who uses it in their business. If you use an AI tool for recruitment, you are a deployer, whether you hire in-house or on behalf of clients.
As the deployer of a high-risk system, you must among other things:
- Use the system in line with the provider's instructions.
- Ensure human oversight. The people overseeing the system must have the competence, training and authority to override it.
- Make sure input data is relevant to the purpose, to the extent you control it, for example the requirements you search on.
- Monitor how the system works and report serious problems to the provider.
- Keep the logs the system generates for at least six months, to the extent they are under your control.
- Inform employees and their representatives before a high-risk system is put into use in the workplace.
- Inform people affected by decisions where a high-risk system is used.
Some deployers, mainly public bodies, must also carry out a fundamental rights impact assessment before they start using the system.
What the provider is responsible for
The provider of an AI system for recruitment must, among other things, have a risk management system, data governance, technical documentation, logging, clear instructions for users, built-in support for human oversight, and adequate accuracy and security. The system must be assessed against the requirements and registered in the EU database before it is placed on the market.
Ask your providers how they work on this. An answer that only says "we comply with the AI Act" is not enough. Ask for concrete examples of how human oversight and explanations work in the tool.
How Essentiq approaches the AI Act
Essentiq is an AI tool for recruitment and is therefore covered by the high-risk requirements from 2 December 2027. We are building the product for them. This already applies today:
- People decide. The AI makes no decisions about candidates. You choose who is saved, contacted and moved forward.
- Explained matches. Every candidate is shown with what is met, partly met and missing against your requirements, so you can review the assessment.
- Labelled AI content. AI-generated analyses and labels are marked as such.
- Servers in Sweden. Servers, database and search index are in Stockholm, and your data is not used to train general AI models.
More on security, the GDPR and the AI Act is on the security page.
The AI Act and the GDPR
The AI Act does not replace the GDPR. All personal data about candidates is still covered by the GDPR, with requirements for a legal basis, information to candidates and their right to object. The GDPR also has rules on decisions based solely on automated processing. Read more in the guide to GDPR in sourcing.
Five things you can do now
- Map your AI tools. Which tools in your hiring use AI, and for what?
- Train the team. AI literacy is already a requirement. Go through what the tools do and where they can get things wrong.
- Decide who has oversight. Who reviews the AI's suggestions, and are they allowed to override them?
- Ask your providers how they are preparing for December 2027.
- Review the information you give candidates, under both the GDPR and the AI Act.
Common questions
Is AI in recruitment high-risk under the AI Act?
Yes. AI systems intended to be used to recruit or select people, for example to target job ads, filter applications or evaluate candidates, are high-risk under Annex III of the AI Act.
When do the high-risk requirements start to apply to recruitment?
After the EU's decision on the Digital Omnibus (Regulation (EU) 2026/1744), the requirements for high-risk AI in Annex III, which includes recruitment, apply from 2 December 2027.
What applies already?
Since 2 February 2025, the bans on certain AI practices apply, including emotion recognition in the workplace, as does the requirement that staff using AI have sufficient AI literacy. Since 2 August 2026, the transparency rules in Article 50 apply.
Who is responsible, the provider or us as users of the tool?
Both, for different things. The provider is responsible for the system meeting the requirements, including risk management, documentation and instructions. You as the user (the deployer, in the regulation's terms) are responsible for how it is used, for example human oversight and informing the people affected.
Does the AI Act replace the GDPR?
No. The GDPR still applies to all personal data about candidates. The AI Act adds requirements on the AI systems themselves and how they are used.
Is this legal advice?
No. This article is general information about the AI Act. Talk to a lawyer for an assessment of your own situation.
By the Essentiq team· Published · Updated
Keep reading
- SecuritySecurity, GDPR and the EU AI ActHow Essentiq protects data: servers in Stockholm, encryption, GDPR for candidates and customers, and how the service is built for the EU AI Act.
- GuideGDPR when sourcing candidates in Sweden: a practical guideHow to source passive candidates in Sweden within GDPR: legitimate interest, informing candidates, data minimisation, retention and their rights.
- EssentiqPrivacy policy: how Essentiq handles dataHow Essentiq AB processes personal data: the profile index, customer data, AI, sub-processors and where they are, retention, security and your rights under GDPR.
- ProductAI candidate search for SwedenDescribe the role in plain language. Essentiq searches over 4 million Swedish profiles, shows the best matches first and explains each one. Free to start.
