Deep Care logo white
Deep Care Logo

EU AI Act: Impact on the use of AI in occupational health prevention

July 5, 2024
The picture shows a woman using the AI assistant Isa at her desk. Above the picture is a transparent EU flag with the inscription

The use of artificial intelligence (AI) has become an integral part of our everyday lives, both at work and in our personal lives. In order to strengthen confidence in this technology and create clear guidelines for its use, the Council of the 27 EU member states adopted the EU Artificial Intelligence Act (AI Act) on 21 May 2024. This regulation marks the world's first comprehensive set of rules for regulating AI and establishes a standardised framework within the European Union.

What is the EU AI Act and what does it mean?

The AI Act establishes clear rules for the development, commercialisation and use of AI systems within the EU. It follows a risk-based approach in which regulation varies depending on the potential risk of an AI application. AI systems are divided into three main categories:

  1. High risk: This usually includes AI systems that are used in sensitive areas such as critical infrastructure, healthcare, education or public safety. Such systems are subject to strict regulations to protect fundamental rights and prevent misuse. Examples include AI systems for biometric identification or credit scoring that process personal data.
  2. Limited risk: AI applications that do not exceed certain risk thresholds must fulfil transparency and information obligations. One example of this is systems that analyse workplace data but do not create comprehensive personal profiles.
  3. Minimal risk: This category includes AI systems that pose little or no risk to the rights and freedoms of users and are therefore subject to minimal regulatory requirements. One example is the digital AI assistant "Isa", which works locally and does not create comprehensive personal profiles.

Further links on the topic:

Effects on EU member states and Germany

For EU countries such as Germany, the AI Act means harmonising regulatory standards for AI, which increases market transparency and facilitates the cross-border use of AI technologies. Germany, as one of the leading economic powers within the EU, sees the clear regulation of AI as an opportunity to both promote innovation and ensure the protection of its citizens. It is not yet clear how exactly the German government intends to transpose the EU's requirements into national law.

Further links on the topic:

Use of AI in occupational prevention

With the exception of the AI assistant Isa few AI systems are currently used in German companies in the context of occupational prevention. In principle, however, AI can be used to better identify risks, derive more targeted measures and personalise or automate health programmes.

AI-based systems can collect and analyse data from various sources, for example to identify employee health risks. As in the case of Isa, this can include data on working conditions or health behaviour, for example - and in the future it will probably even include mental health parameters . Based on this data, companies can take preventative measures to promote health and safety in the workplace. AI can therefore recognise potential hazards at an early stage in order to derive targeted preventive measures or initiate them itself. It can also be used to create personalised health programmes for employees based on their individual needs and health goals. Automated health advice systems can offer employees continuous support by responding to their personal health data and behavioural patterns and making appropriate recommendations.

How will the EU AI Act and the subsequent regulations from the German government affect AI applications in occupational health management?

The AI Act provides clear guidelines for providers, users and decision-makers alike:

  • Transparency and information obligations: Providers of AI-based solutions in occupational safety and health management must ensure that users are fully informed about how the systems work and how their data is used.
  • Data protection and security measures: Compliance with data protection standards and the implementation of appropriate security measures will be monitored more closely to ensure the protection of personal data.

This gives companies and employees who request or use AI offerings more extensive information rights with regard to AI systems and the use of their data. At the same time, the likelihood of data misuse is reduced, which strengthens consumer confidence in AI-based products and services and promotes the acceptance of such technologies.

Classification of "Isa" under the EU AI Act

Our digital AI assistant Isa is categorised as "minimal risk" under the AI Act due to its locally operating and data protection-compliant nature. This means that Isa is only subject to minor regulatory requirements, which facilitates its acceptance and implementation in occupational health management, occupational health and safety and health insurance schemes.

Another reason for this categorisation lies in How Isa works. Even in the event of a malfunction or a wrong decision, Isa poses no immediate danger to users. Isa gives users recommendations to improve ergonomics or to drink enough and teaches skills for healthy habits. However, if Isa's analyses had an influence on a user's medical treatment, this would lead to a higher risk classification.

A clear framework for a secure and innovative future

The EU AI Act represents a pioneering step towards promoting the potential of artificial intelligence in Europe without jeopardising the fundamental rights of citizens. It creates a clear framework for the use of AI technologies and at the same time offers the flexibility to drive innovation. Germany and other EU countries face the challenge of integrating the AI Act into their national legal systems in order to ensure the safe and responsible use of artificial intelligence. AI companies will be held more accountable for ensuring the safety of their systems, creating more transparency and complying with clear legal requirements. Provided that the set of rules is well implemented in EU countries, it can create a level playing field for companies that want to achieve high data protection and security standards for their AI systems and adhere to ethical principles.

Stay up to date