Navigating the EU AI Act
Understand your obligations under the world's first comprehensive AI law. We help you stay compliant with High-Risk AI system regulations.
The Risk-Based Approach
The EU AI Act categorizes AI systems into four levels of risk. Most AI used in employment (CV scanning, candidate ranking) falls into the High-Risk category.
- Unacceptable Risk (Prohibited)Emotion recognition in workplaces, social scoring. Banned completely.
- High Risk (Regulated)Recruitment, CV filtering, promotion decisions. Subject to strict compliance.
- Limited Risk (Transparency)Chatbots, deepfakes. Users must know they are interacting with AI.
Check Your Risk Classification
Determine where your hiring technology falls under the EU AI Act.
Is Your Hiring AI "High-Risk"?(EU Regulation 2024/1689)
The EU AI Act classifies AI systems by risk. Most employment and HR systems fall into the "High-Risk" category.
Key Risk Categories Checked:
- Prohibited: Unacceptable risk (e.g., emotion recognition in workplaces)
- High-Risk: Recruitment & candidate evaluation tools
- Transparency: Chatbots and user interaction
Obligations for High-Risk Systems
1. Data Governance
- • Training, validation, and testing data must be relevant and representative
- • Check for biases and mitigate them
2. Technical Docs
- • Maintain detailed technical documentation
- • Demonstrate how the system works and how risks are managed
3. Human Oversight
- • Designed to be overseen by natural persons
- • Ability to stop or override the system
- • "Human-in-the-loop" is mandatory
Provider vs. Deployer
The Act distinguishes between the Provider (who develops the AI) and the Deployer/User (who uses it for hiring).
| Responsibility | Expert Hire (Provider) | You (Deployer) |
|---|---|---|
| CE Marking & Conformity Assessment | ||
| Risk Management System | ||
| Data Governance & Bias Mitigation | ||
| Ensure human oversight during use | ||
| Follow instructions of use | ||
| Inform workers/candidates (Transparency) |
We Handle the Heavy Lifting
As the provider of the AI system, Expert Hire carries the provider-side obligations: conformity assessment, quality management and logging. The deployer obligations stay with you as the employer, and no vendor can take those on for you.
Automatic Logging
The system automatically logs all outputs, conformities, and risk assessments for future audits.
EU-Specific Controls
EU and UK data residency is available on request and provisioned for you before rollout. Practices Article 5 prohibits, such as emotion recognition, are not offered to anyone: we do not score facial expression, tone or affect in any market.
Frequently Asked Questions
When does the EU AI Act apply?
The Act entered into force in mid-2024, with full application of high-risk rules phasing in over 24-36 months. However, early compliance is strongly recommended.
What are the fines?
Fines can be up to €35 million or 7% of global turnover for prohibited practices, and up to €15 million or 3% for other non-compliance.
I am a US company. Does this apply to me?
Yes, if you place systems on the EU market or if the output of your system is used in the EU (e.g., assessing an EU candidate).
Future-Proof Your Hiring
Adopt a system built for the strictest global standards from day one.
Create Free Account