Recruiting and the EU AI Act: Navigating the New Rules

The EU AI Act is poised to reshape the recruitment landscape, placing guardrails on the use of Artificial Intelligence (AI) to ensure fairness and protect candidate rights. While the Act isn't fully implemented until late 2025, some key provisions, particularly those concerning prohibited AI practices, are already provisionally applicable.

This article focuses on what's allowed and what's not allowed in AI-driven recruitment under the evolving framework of the EU AI Act, specifically considering the impact of provisions already in effect and those coming soon.

Disclaimer: This article is for informational purposes only and does not constitute legal advice. The EU AI Act is a complex and evolving piece of legislation. It is essential to conduct your own thorough research and consult with legal counsel specialising in EU law and AI for guidance tailored to your specific circumstances. Consider employing an internal or external compliance team experienced in EU regulations to ensure your organisation's AI practices align with the Act's requirements.

The Shifting Sands: What's Changed?

The EU AI Act introduces a risk-based approach, classifying AI systems based on their potential impact. Recruitment AI typically falls under the "high-risk" category, subjecting it to stricter scrutiny. While AI offers significant potential for streamlining and optimizing recruitment, the Act sets clear boundaries to prevent misuse. It's crucial to understand that while many provisions come into force late 2025, some are already in effect, focusing on outright bans of particularly harmful AI practices.

What's Allowed (with caveats):

AI can still be a valuable tool in recruitment, but its application must align with the principles of fairness, transparency, and human oversight. Here are some examples of what's generally permissible, but always with the caveat that compliance with the full Act is essential:

• Automating Administrative Tasks: AI can streamline repetitive tasks like scheduling interviews, sending reminders, and managing applications. This frees up recruiters to focus on more strategic activities.

  
  

• Analysing Large Datasets (with caution): AI can analyse large datasets of resumes or applications to identify trends, skills in demand, and potential talent pools. However, this analysis must be carefully monitored to avoid perpetuating existing biases. Data quality is paramount, and techniques like explainable AI (XAI) should be considered.

  
  

• Personalised Job Recommendations: AI can match candidates with suitable job openings based on their skills and experience. However, these recommendations must be based on objective criteria and avoid reinforcing discriminatory patterns. Transparency in the recommendation process is key.

  
  

• Initial Screening (with human oversight): AI can assist in initial screening of resumes or applications to filter out clearly unqualified candidates. However, this screening must be transparent, and human review is essential to prevent errors or biases from the AI system. Clear guidelines for human intervention are necessary.

  
  

What's NOT Allowed (or highly restricted):

The EU AI Act prohibits or severely restricts certain AI applications in recruitment due to their potential for harm:

• Biometric Categorisation: AI systems that use biometric data (facial recognition, emotion analysis, etc.) to categorise candidates based on protected characteristics like race, ethnicity, gender, sexual orientation, or disability are strictly forbidden. This is a core principle of the Act, designed to prevent discriminatory profiling.

  
  

• Manipulation or Exploitation: AI systems that are designed to manipulate candidates' behaviour or or exploit their vulnerabilities are also banned. This includes AI that attempts to subtly influence a candidate's responses during an interview through persuasive language analysis.

  
  

• Discriminatory Profiling: AI cannot be used to create profiles of candidates based on protected characteristics and then use those profiles to make hiring decisions. Even if the AI doesn't explicitly categorise based on race or gender, if its analysis leads to discriminatory outcomes, it's a violation.

  
  

• Automated Decision-Making without Human Oversight:  While AI can assist in various stages of recruitment, it cannot be used to make final hiring decisions without meaningful human review. Human oversight is essential to ensure fairness and prevent AI errors from having disproportionate impacts. Clear processes for human intervention are crucial.

  
  

Navigating the Grey Areas:

The EU AI Act also presents some grey areas that require careful consideration:

• Implicit Bias in Algorithms: Even if an AI system doesn't explicitly discriminate, it can still perpetuate existing biases if it's trained on biased data. Addressing this requires careful data analysis, ongoing monitoring of AI outputs, and potentially techniques like adversarial debiasing.

  
  

• Explainability of AI Decisions:  Candidates have a right to understand how AI has been used in evaluating their applications. Recruiters must be able to explain the factors that influenced AI-driven decisions, which can be challenging with complex algorithms. Explainable AI (XAI) is a critical area of development.

  
  

Preparing for Compliance:

While the full Act comes into force later, the time to prepare is now. Recruiters should:

• Conduct AI Audits:  Evaluate current AI tools for compliance with the EU AI Act.

• Prioritise Data Quality:  Ensure data used for training AI is high-quality, unbiased, and compliant with GDPR.

• Invest in Transparency:  Develop strategies for explaining AI-driven decisions to candidates.

• Train HR Staff:  Educate HR teams on the requirements of the AI Act and how to use AI responsibly.

• Seek Legal Counsel:  Consult with legal experts specialising in AI and data privacy.

The EU AI Act significantly impacts the use of AI in recruitment.

While AI offers valuable tools for efficiency and insight, its deployment must be carefully managed to ensure fairness and protect candidate rights.

By understanding the allowed and restricted practices and proactively preparing for compliance, recruiters can navigate this new regulatory landscape and leverage AI responsibly.

Sources:

• European Commission's website on the AI Act

• Council of the European Union's press release on the AI Act: (Search for the most recent press release on the Council's website related to the AI Act)

• European Parliament's website on the AI Act: (Search for relevant documents on the Parliament's website)

  
  

Remember to consult official EU sources and legal counsel for the most up-to-date and accurate information.

This information is for guidance only and should not be considered legal advice.

About the Author

Clarice, Founder of The Talent Seed

Clarice is a seasoned HR leader with a proven track record of scaling teams from 20 to 5000+ employees. With over 18 years of experience across Asia, Australia, and Europe, she's built high-performing teams for industry leaders (Shell & PwC) and FinTech (N26, Billie & Solaris) startups.

As a certified career coach, Clarice helps individuals reach their full potential, guides leaders and organisations, attracts top talent, enhances employee retention, and navigates the evolving talent market.

The Talent Seed is a cutting-edge recruitment consultancy revolutionising the way companies hire by combining the power of human connection with advanced recruiting technology. Dedicated to empowering hiring leaders to navigate growth with confidence. Contact The Talent Seed today and discover how we can help you unlock the full potential of your talent acquisition efforts.

Join us in our mission to elevate the recruitment process across the globe.

Follow us on our LinkedIn & Instagram pages for exclusive updates, hiring tips, recruitment technology trends and expert discussions.