In the latest instalment of our special series marking 30 years of HÄRTING, we are joined by a familiar face: Marlene Schreiber, partner in the IT team and Head of AI at HÄRTING. Together with Dr Martin Schirmbacher and Carl Nowak, she discusses how companies can use artificial intelligence in a legally compliant manner and why the AI Act covers only part of the legal challenges involved.
The starting point is the AI Officer training programme that Marlene and Martin offer in collaboration with Blackboat. After all, anyone using AI within a company must take far more into account than just the AI Regulation: data protection, copyright, employment law, liability issues and effective AI governance also play a key role.
The episode focuses on two particularly practical topics: high-risk AI and transparency obligations under the AI Act. When is an AI system classified as high-risk? What specific considerations apply, for example, when using it in recruitment processes or in the education sector? And what organisational measures must companies put in place?
The episode also addresses the labelling of AI-generated content. Not every piece of AI-generated text needs to be labelled. However, special transparency requirements apply to deepfakes and certain types of synthetic audio and video content. Using examples from radio advertising and e-commerce, the three discuss how sensible and practicable these obligations are.
An episode on the challenges of AI regulation, unresolved legal issues and the practical implementation of the AI Act within companies.