What the EU AI Act really means for interpreting.

In this new AIIC Science Hub conversation, Irina Paramonova and Pavol Šveda sit down with Professor Yaniv Benhamou, Director of the Digital Law Center (University of Geneva), to unpack what the EU AI Act really means for interpreting. Drawing on his work on creative value chains and platforms, Yaniv explains how AI is reshaping “intellectual professions” – from artists and lawyers to interpreters – and how value is increasingly captured by platforms trained on our work. He then walks us through the EU AI Act as a product-safety style, risk-based framework, clarifying where automated speech translation (AST) and CAI tools fit, and when their use in justice, migration, healthcare or HR could qualify as high-risk, thereby triggering special legal safeguards. Key themes include the distinction between providers and deployers, transparency duties, and the requirement for meaningful human oversight in real time, including a clear “fallback to human” mechanism when AI output becomes unreliable. The discussion also touches on GDPR, consent and data governance when client speech and interpreters’ voices are used to train models, the emerging licensing market for training data, and first national experiments such as the Italian AI law. Throughout, Yaniv offers a cautiously optimistic vision of AI as an assistive tool, not a replacement, and points to concrete levers interpreters and their associations can use in shaping this future.