Artificial intelligence has long since arrived in our everyday lives. Quickly formulating a rejection to an applicant with ChatGPT here, creating a personal training plan for an employee via AI there and testing the new AI functions in the HR software in between. The challenge is that there is still a lot of legal uncertainty surrounding the use of AI. The EU’s AI Act is intended to provide a remedy. But what exactly does it actually regulate?
One in five companies in Germany now utilises the benefits of AI. That is eight percentage points more than a year ago, almost doubling from 12 to 20 per cent. The rapid spread of AI also makes it important for the EU to regulate the technology – we now all know what economic and ethical damage can otherwise be caused by data breaches, deep fakes and the ‘improper’ use of AI. To prevent this, there is the EU AI Act, often referred to simply as the AI Act in Germany.
To what extent is AI being used in HR in Germany?
Before we explain in detail what the current status of the implementation of the law is and what it means for you, let’s take a look at how AI is actually being used in German HR offices and personnel departments. Because: 20 per cent of companies use AI – that’s clear. But how exactly?
First of all: statistics on the use of AI in HR – the 20 per cent figure is separate from specific areas of the company – vary widely.
The ifo-Institute calculated a share of five per cent for Germany at the end of 2023, while a survey by HR consultancy Heidrick & Struggles puts the figure at 87 per cent for Europe.
We are certainly not that far behind in Germany – even if other countries are often one step ahead of us here. What should be clear, however, is that all studies show the same thing: More and more AI applications are also finding their way onto the digital table in HR departments.
The areas of application are diverse:
- Automated applicant selection: AI filters CVs and application letters to quickly identify suitable talent.
- Chatbots in Recruiting: AI chatbots answer applicant questions and conduct initial interviews – according to Heidrick & Struggles, these first two points are the primary areas of application for AI in HR.
- Predictive analytics for talent management: AI analyses data to identify employees’ potential and development opportunities.
- Automation of administrative HR tasks: Routine tasks such as contract documents and holiday calculations are taken over by AI tools.
- Personalised further training:AI creates customised learning plans based on individual skills and potential.
- Employee loyalty and satisfaction: AI tools analyse feedback to identify risks of employee turnover at an early stage – this is already possible with ChatGPT.
The areas of application are therefore diverse, ranging from the automation of administrative tasks to support in the recruitment process and personalised employee development. And this is where problems lurk – Amazon, for example, used AI in recruitment to select applicants and the system disadvantaged female applicants. In other words, clear rules, guidelines and regulations are needed here, otherwise the sustainable use of AI will not work.
EU AI Act: The current status of the AI law
This is exactly where the EU AI Act comes in. The AI Act was already proposed in an initial version in 2021 and came into force on 1. August 2024. For member states such as Germany, this means that the first implementations must be completed by 2 August 2025.
The legal text categorises AI systems according to risk levels, whereby applications in the HR sector in particular are often classified as ‘high-risk AI’ – this applies to AI-supported systems for applicant selection, performance evaluation or personnel development, for example. Chatbots, for example, are categorised as AI with limited risk, but as a company you must also observe transparency obligations here – applicants must be informed that communication is taking place with an AI, for example.
Strict requirements apply to high-risk AI systems:
- Risk management: Companies must implement systematic risk management in order to identify potential risks at an early stage and take countermeasures.
- Transparency and documentation: The EU AI Act requires comprehensive documentation of the AI systems used – this also includes the data sources used and decision-making processes in the background.
- Human supervision: The systems must be designed in such a way that human specialists can monitor the decisions and intervene if necessary.
In short, the EU AI Act will sooner or later force you and your HR department to scrutinise AI systems and their use and ensure that they comply with both legal requirements and ethical standards. IT managers in particular are therefore required to have clear expertise in this area.
In addition, companies must not only make the pure use of AI transparent – it must also be ensured that the technology is used fairly and without discrimination in human resources in order to prevent a second Amazon. The following applies to applicant selection and performance assessment in particular: there must be no unintentional distortions or discrimination by the systems used, and this is particularly influenced by the training data of the systems.
These increased requirements under the AI Act entail many tasks – for everyone working in HR. This also means that HR employees will have to mark a few more training days in their diaries in future to ensure that the use of AI tools also fulfils the new legal requirements.
EU AI Act summarised: What you need to watch out for
Briefly summarised: The EU AI Act is a law that came into force on 1 August 2024 to regulate artificial intelligence in the EU. It is intended to make AI applications safe and ethical – which means for HR departments and HR managers:
- AI systems are categorised according to risk levels, with HR applications that support the selection of applicants, for example, falling into the high-risk category.
- Companies must implement appropriate AI risk management in order to minimise potential risks. The data sources and decision-making processes used in the background.
- Transparency obligations require detailed documentation of the data used and decision-making processes in the background.
- All AI systems must be under human supervision to prevent discriminatory decisions.
- The timely implementation of these requirements is necessary by 2 August 2025.
Conclusion: More rules, but also more opportunities?
A glance at the new regulations may cause a blockade at first – more rules, more work. Does this mean the end of AI systems in your company? The answer is a clear no. After all, AI offers unimagined potential for increasing the efficiency of HR processes.
Ki-Assistant Rai in the rexx Suite, for example, acts like a personal assistant who supports the user with questions and helps to solve challenges in order to get the best out of the rexx Suite.
Despite rules and regulations, there is no alternative for companies to deal with AI in HR in order to stay ahead of the competition. So see it less as an imposed duty to take care of the implementation of and compliance with new AI guidelines – and more as an investment in your company’s future.
You might also be interested in:
