Young people with smartphonesdts
.
Berlin (German news agency) – The IT industry association Bitkom considers the debate initiated by Federal Minister of the Interior Faeser (SPD) about mandatory labeling for AI-generated content to be premature. First of all, society must be informed about what artificial intelligence means and what it can achieve, said Lukas Klingholz, Head of AI and Cloud at Bitkom, the dts news agency.
“Almost half of Germans don’t know what artificial intelligence actually is.” Therefore, it must first be clarified about AI, what it can and cannot do and how artificial intelligences differ from each other. “Without this basic knowledge, a labeling obligation would be useless,” says the expert. Another challenge lies in the fact that many companies use the AI term to upgrade their offerings to customers: “It’s still too often the case that AI isn’t in everything that says AI.”
In general, Klingholz considers a labeling requirement to be technically feasible, and some companies are already offering solutions: “It can be a digital watermark or an artificial intelligence that indicates the probability of whether a text was written by another AI.” The quality still varies, but the Bitkom expert is counting on technical innovations that will make such labeling possible in the coming years. Many companies are already proactive in this regard, since transparency and trustworthiness are indispensable for platforms and media as well as for AI-based systems: “Therefore, this development should be observed and evaluated before you follow the path of legal requirements, the concrete form of which is large would bring challenges,” said Klingholz to the dts news agency.
HOME PAGE