Based on the AI Act, the European Commission wants to more strongly regulate artificial intelligence in the future. Tomorrow, Thursday, the lead committees IMCO and LIBE will vote on the Commission’s draft report.
eco – Association of the Internet Industry advocates a risk-based approach in the AI Act, meaning that the requirements for AI systems would be proportionate to the risk in the respective areas of application.
For instance, AI systems that are used to protect critical infrastructure or the area of law enforcement should be subject to higher requirements than systems such as those used in the area of customer service to process enquiries.
In addition, the eco Association advocates the prohibition of certain practices, such as biometric surveillance in public spaces or state scoring of citizens.
At the same time, eco warns against extending the AI regulation to systems without a specific field of application – so-called general-purpose AI, which also includes language models such as ChatGPT.
eco Chair of the Board Oliver Süme has the following to say:
“It is naturally important that the AI Act clearly addresses potential risks, but in order to harness the opportunities of the technology, we need room for innovation. Today, Europe is already lagging behind in the development of AI systems. This threatens not only Europe’s technological sovereignty, but also the competitiveness of the entire economy in Germany.
“The AI revolution is only just beginning. I think it is especially important now to develop and train AI systems on the basis of European values so that they can set standards worldwide. This also includes enabling real labs to give developers the chance to recognise malfunctions on time.”
In January of this year, the Feasibility Study of the Large European AI Models (LEAM) initiative launched by the German AI Association also highlighted that AI will enable many new applications, platforms and business models in the future. The eco Association is one of the co-authors of the study.
Link to the topic of AI
