It’s getting scary: Machines learn to draw pictures, imitate voices or replicate facial expressions. This raises new questions for society in dealing with so-called artificial intelligence (AI). How does this supposedly creative power of non-human creation fit into our value and legal system? How do we want to deal with this in terms of copyright? What data protection issues could we face in the near future?
Andrea Trinkwalder and Jo Bager podcasting
In the current episode of the c’t data protection podcast, Heise lawyer Joerg Heidrich and c’t editor Holger Bleich discuss this. They are supported by Andrea Trinkwalder and Jo Bager, who have dealt intensively with the topic in the c’t editorial team. The two first introduce the topic and explain how the development leap in AI came about. Trinkwalder explains in an easy-to-understand way how the currently popular AI image generators work.
Heidrich then outlines the legally affected problem areas and warns against scaremongering: Yes, data protection plays a role. But mostly data – also for training purposes – would be processed, which are already published. In the case of use, the “legitimate interest” should prevail in a consideration. The prerequisite is that the AI products sufficiently alienate the data fed in. The copyright aspects are trickier. Here the discussion is in full swing, as Heidrich, Trinkwalder and Bager show with examples.
Episode 75:
Here are all the episodes so far:
(raised)