The technical inspection association TÜV has said that the implementation of the EU Artificial Intelligence Act at national level in Germany should not be put off until the next government is formed.

According to TÜV Managing Director Joachim Bühler, the law forms the basis for ensuring that Europe and Germany do not fall behind in the competition for AI technology. In view of the collapse of the coalition government, delays in this field would be „harmful in terms of competition policy“, he said at the presentation of a new study on the use of AI in Germany.

„We now need to quickly develop a standardised leading market for AI testing and certification, so that trustworthy AI – made in Europe – becomes a unique selling point,“ says Bühler. In order to be able to regulate real AI risks in a targeted manner, it is also important to establish systematic AI damage statistics. Furthermore, the implementation of the EU AI Act is necessary in order to meet requirements for transparency, risk management and labelling of AI-generated content.

In the study on the second anniversary of the presentation of the ChatGPT AI model, it became clear that a significant majority of Germans demand transparency and labelling obligations for AI-generated content (90%). In addition, Germans consider mandatory safety testing of AI systems by independent testing organisations to be a prerequisite for products and applications with AI to be safe and ethical (83%).

This is based on the concerns of 87% of respondents who fear that they will not recognise fake images and videos (deepfakes) made with the help of AI. 77% also state that they are unable to check the truthfulness of news on the internet („fake news“). This would provide scope for influencing elections, and harm democracy and public discourse.

In this context, TÜV Managing Director Bühler is calling for „clear guidelines for the use of AI in election campaigns, and voluntary commitments by political parties to use AI responsibly“. The operators of social media platforms would have to take concrete measures to protect democratic processes. This is why the draft bill for the German implementation of the EU AI Act must be finalised quickly. This would create „not only planning and legal certainty for companies and testing organisations, but also a higher level of protection for citizens“.

Overall, the survey of around 1,000 people aged 16 and over shows that every second German citizen (53%) now uses generative AI. In October last year, this figure was only 37%. „AI services are developing into indispensable everyday digital helpers similar to search engines, emails or navigation services in both the private and professional spheres,“ says Bühler.

A clear majority agree with the statement that AI technology has the potential to support them in their private and professional lives. However, a clear majority also want to be protected from the potential risks (76%). And as AI is increasingly developing into a work tool, the accuracy of the results is all the more important. However, half of users currently have little or no trust in AI results. Appropriate regulation is therefore even more important to increase acceptance, warns TÜV.