After almost three years of negotiations, the relevant EU institutions have agreed on a final, oral text for the AI Regulation. This is due to come into force in April 2024 at the latest, with the first provisions already applying in October of the same year. Designed as a comprehensive set of rules, the AI Regulation addresses aspects of the production, development and use of AI systems.
The regulation categorizes AI systems into four risk levels. The lowest level comprises systems with minimal risks, such as recommendation algorithms, which do not entail any further obligations. The next highest category includes systems with “specific transparency obligations”, including chatbots. Although these can potentially spread misinformation, they are not classified as dangerous.
The third level includes high-risk systems, such as AI in critical infrastructure or credit checks, which are subject to stricter regulations. The fourth and final level prohibits certain applications in principle, including social scoring, facial data scraping and emotion recognition.
Companies could be exposed to a higher financial risk, as software is now considered a product or product component for product liability purposes. There will be no specific AI-related liability directive before the European elections in June 2024. To avoid delays, this is integrated into the general product liability directive. Fines for breaches of the AI Regulation can amount to up to 7% of a company’s global annual turnover.
Conclusion: The AI Regulation will have a significant impact on the future in Europe and therefore represents an important milestone for companies. As an IT consultancy specializing in the introduction of AI solutions, we and our clients are facing new challenges and opportunities. This regulation requires us to adapt our strategies and processes to ensure compliance while delivering innovative solutions.
EU Internal Market Commissioner Thierry Breton emphasized the importance of the regulation: “The AI law is much more than a set of rules – it is a launch pad for start-ups and researchers in the EU to take the lead in the global race for trustworthy AI”. This underlines the importance of the regulation for companies that want to be at the forefront of technological innovation.
More links on the topic on Sulzer.de:
Turn enterprise data into competitive advantage with AI – Sulzer GmbH
ChatGPT and data protection – How strict are OpenAI and Microsoft?
Contact us for customized advice:
To learn more about how Sulzer GmbH can help your company navigate the new AI regulations and to discuss customized solutions, contact us. We look forward to working with you to shape the future of AI in Europe.