AI Act: what are the stakes for HR?

Published in

Future of Work & Society — Towards the quaternary economy

8 min readJun 26, 2024

Artificial Intelligence (AI) is set to revolutionize the world of work, and the HR function is no exception. With the recent adoption of the AI Act by the European Union, the first comprehensive regulatory framework on AI worldwide, companies will have to adapt. What are the concrete challenges and impacts of this legislation for HR? How can it be leveraged to improve HR processes while remaining ethical and compliant? Let’s break it down.

AI Act: anecessary framework for ethical and responsible AI

Let’s start from the beginning: the AI Act was proposed by the European Commission in April 2021 and was definitively adopted on May 21, 2024. It is the first comprehensive regulation on AI worldwide, aiming to harmonize the rules for the placing on the market, putting into service, and use of AI systems in the European Union. The adoption process included numerous debates and consultations, highlighting the importance of this regulation (European Commission, 2024).

We can mention three main objectives for this initiative:

Protection of fundamental rights: in line with the GDPR, ensuring that AI systems respect the fundamental rights and values of the European Union.
Promotion of innovation: supporting innovation while ensuring an ethical use of AI.
Creation of a harmonized legal framework: establishing clear and harmonized rules to avoid fragmentation of the internal market (European Commission, 2021).

Indeed, AI raises many ethical questions, particularly regarding respect for fundamental rights, transparency of algorithms, or risks of discrimination. With nearly a quarter of employees already using AI, often without their employer’s knowledge (IFOP/LearnThings, 2024), a framework was necessary. The AI Act thus stands out from other international regulations, such as those of the United States or China, by placing particular emphasis on the protection of fundamental rights and ethics. For example, while the United States adopts a more innovation and competitiveness-focused approach, the EU’s AI Act focuses on protecting citizens’ rights and transparency.

The AI Act aims to establish a harmonized framework for the development and use of AI in the EU. Its risk-based approach imposes graduated requirements according to the risk level of AI systems:

Unacceptable risk (e.g., social scoring systems): outright prohibition
High risk (e.g., recruitment, evaluation, workplace surveillance): strict requirements for transparency, traceability, human oversight, and risk assessment
Limited risk (e.g., chatbots): transparency obligations and user information
Minimal risk (e.g., spam filters): no obligations

AI Act: implications for the HR Function

AI is changing the organization and execution of work. Companies must regulate the use of AI by employees, for example by integrating specific rules into the internal regulations or IT charter. An IFOP study for LearnThings (2024) revealed that 22% of employees have already used an AI tool in a professional context, 55% of them without informing their manager. A study conducted by Deloitte (2023) shows that 45% of European companies have already reorganized their work processes to integrate AI. And even for HR processes, AI is going to change the game.

Recruitment

Recruitment is one of the most mature HR areas in terms of AI adoption, but also one of the most sensitive. CV screening, chatbots, automated video interviews, predictive tests… AI promises faster and more efficient processes, and also very biased ones! Amazon learned this the hard way in 2018 with its CV pre-selection tool, trained on mostly male profiles and reproducing this bias. Since then, many vendors like Textio or TapRecruit offer de-biasing and data anonymization solutions.

The AI Act classifies these systems as high-risk and subject to enhanced requirements:

Transparency on evaluation criteria and methods;
De-biasing and representativeness of training data;
Possibility of appeal and human intervention;
Regular risk assessment and corrective measures;
Certification and registration with authorities.

Companies must ensure that their AI recruitment tools comply with the requirements of the AI Act to avoid fines of up to 35 million euros or 7% of annual worldwide turnover (European Commission, 2024). A real paradigm shift that will force companies and vendors to review their practices. But also an opportunity to make recruitment processes fairer and more diverse, because it’s not as if everything was rosy in this area.

Performance evaluation

Another sensitive area is the evaluation of employee performance and surveillance by AI. While some see it as a way to objectify evaluations, others fear dehumanization and over-quantification of work. The AI Act sets clear limits:

Prohibition of AI systems inferring emotions or psychological state of employees, except for exceptions (security, research);
Prohibition of AI systems evaluating sensitive personal characteristics (ethnicity, political or sexual orientation, etc.);
Obligation to inform employees about the use of AI to evaluate or monitor them;
Right to challenge decisions made by AI and obtain human intervention.

Microsoft had to revise its “Performance AI” system following employee concerns. Now, employees have more control and transparency over the data and criteria used. The challenge for HR will be to find the right balance between the potential of AI (more frequent and targeted evaluations, detection of psychosocial risks, etc.) and respect for employees. A cultural change that will require a lot of education.

Learning & development

According to the World Economic Forum, 50% of employees will need reskilling by 2025 due to automation and AI. A huge employability and upskilling challenge for HR. But AI can also be a valuable ally, by personalizing training paths. Platforms like Coursera, EdX or the recent Google AI Academy use AI to recommend content tailored to each learner, at their own pace.

Tomorrow, training will be increasingly immersive and interactive thanks to generative AI (virtual reality, role-playing, personalized coaching, etc.). This will make learning more motivating and effective. The AI Act encourages these educational innovations, while regulating the use of learning data. Companies will have to obtain the explicit consent of employees and guarantee them a right of access and rectification. Training in AI itself will be a central issue to demystify the technology, prevent misuse and foster a responsible AI culture at all levels of the company.

According to an IMF report (2023), 60% of jobs in advanced countries will be impacted by AI. According to Gartner (2023), AI could create 500 million jobs by 2033, but these jobs will require specific skills that workers need to acquire. Companies must anticipate these changes and secure career paths in heavily impacted sectors.

Prevention of mental health risks

The use of AI can lead to new psychosocial risks, such as excessive dependence on technology, isolation, work complexification, and loss of meaning. Companies must take measures to limit these risks and ensure the well-being of their employees (European Agency for Safety and Health at Work, 2023). Proposing concrete measures, such as instituting regular breaks and disconnection sessions, can help reduce dependence on technology (Smith & Anderson, 2018). But the subject is much broader, as it touches on work organization and the very definition of what working will mean in the years to come…

Conclusion

AI will deeply transform the HR function. With the AI Act, Europe is setting a framework for this transformation to take place in an ethical and responsible manner. For companies, it is an opportunity to rethink their HR processes by placing people at the center. This will require more transparency, explainability and collaboration with employees in the design and deployment of AI solutions.

HR will have to develop a real AI culture, by training employees but also by surrounding themselves with experts to audit their practices. A new profession is emerging: that of “HR AI Ethics Manager”. Because beyond legal compliance, the challenge is to make AI a lever for engagement, inclusion and talent development. Technology is only a means, the human must remain the end goal.

Note: THANKS to Capstan Avocats, whose legal expertise and recent comments allowed me to consolidate this article.

Glossary

AI (Artificial Intelligence): technology whose objective is to reproduce human capabilities to learn, think and interact.
AI Act: European regulation on artificial intelligence adopted in 2024.
GDPR (General Data Protection Regulation): European regulation on the protection of personal data.
Algorithmic bias: tendency of AI systems to reproduce or amplify biases present in the data used for their training.
High-risk AI systems: AI systems identified by the AI Act as presenting high risks to the rights and safety of individuals.

Bibliography

Agence Européenne pour la Santé et la Sécurité au Travail. (2023). Impact de l’IA sur la santé au travail. Retrieved from https://osha.europa.eu
Binns, R. (2018). Fairness in Machine Learning: Lessons from Political Philosophy. Proceedings of the 2018 Conference on Fairness, Accountability, and Transparency (FAT* ‘18). https://doi.org/10.1145/3287560.3287583
Bogen, M., & Rieke, A. (2018). Help Wanted: An Examination of Hiring Algorithms, Equity, and Bias. Upturn. Retrieved from https://www.upturn.org
Brynjolfsson, E., & McAfee, A. (2017). The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. W. W. Norton & Company.
Capstan Avocats. (2024). Les relations de travail face aux enjeux de l’intelligence artificielle.
Code du travail, art. L2312–8, II 4°.
Deloitte. (2023). The Future of Work: How Artificial Intelligence is Transforming the Workplace. Retrieved from https://www2.deloitte.com
European Commission. (2021). Proposal for a Regulation laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Retrieved from https://ec.europa.eu
European Commission. (2024). AI Act. Retrieved from https://ec.europa.eu
FMI. (2023). The Future of Jobs Report. Retrieved from https://www.imf.org
Gartner. (2023). AI and the Future of Work. Retrieved from https://www.gartner.com
Goodman, B., & Flaxman, S. (2017). European Union regulations on algorithmic decision-making and a “right to explanation”. AI Magazine, 38(3), 50–57. https://doi.org/10.1609/aimag.v38i3.2741
IEEE. (2023). Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Artificial Intelligence and Autonomous Systems. Retrieved from https://ethicsinaction.ieee.org
IFOP. (2024). Étude sur l’utilisation de l’IA au travail. Retrieved from https://www.ifop.com
Orange. (2023). How Orange Uses AI to Manage Talent. Retrieved from https://www.orange.com
Raghavan, M., Barocas, S., Kleinberg, J., & Levy, K. (2020). Mitigating Bias in Algorithmic Hiring: Evaluating Claims and Practices. Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency (FAT* ‘20). https://doi.org/10.1145/3287560.3287587
Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why Should I Trust You?”: Explaining the Predictions of Any Classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD ‘16). https://doi.org/10.1145/2939672.2939778
Smith, A., & Anderson, J. (2018). AI, Robotics, and the Future of Jobs. Pew Research Center. Retrieved from https://www.pewresearch.org
Voigt, P., & Von dem Bussche, A. (2017). The EU General Data Protection Regulation (GDPR): A Practical Guide. Springer.

— —

[Article created on 26 june, 2024, by Jeremy Lamri with the support of the GPT-4 algorithm from Open AI for structuring, enriching, and illustrating. The writing is primarily my own, as are most of the ideas in this article].

— —

Follow my news with Linktree

If you are interested in the combination of web 3 and HR, I invite you to subscribe to the dedicated newsletter that I keep writing on the subject, and to read the articles that I have written on the topic: