ARD is introducing AI voices for parts of its radio offering. Starting March 3rd, traffic and weather reports in the joint programs “Pop – Die Abendshow” and “Popnacht” will be presented by an AI. Acute danger alerts will continue to be handled by the live editorial team.
The evening and night programs are broadcast on the public service stations hr3, rbb 88.8, MDR JUMP, NDR 2, Bremen Vier, SR 1, SWR3, and WDR 2. The two joint programs were created at the beginning of 2025 and have so far broadcast nationally uniform traffic and weather reports. The AI is now intended to enable regionally differentiated reports depending on the broadcasting area. The sound design of the individual stations, as well as regional news, have been individual from the start. SWR3 produces the joint programs.
The AI will only read out texts that have been written and checked by the respective editorial teams. It is not allowed to formulate its own reports or change written content. ARD states that the system is not intended to replace human employees.
ARD also explains that reports voiced by AI will only be broadcast in combination with a corresponding transparency notice. The AI voices are based on those of the real moderation team. “The AI-supported data processing and speech output system was developed and configured in close cooperation between SWR and WDR,” ARD announced. Data protection and security are paramount.
At heise, an AI-generated voice of moderator Isabel Grünewald has been in use in the podcast “Kurz informiert” since 2022.
Share of AI in radio is increasing
At the beginning of the year, ARD, ZDF, Deutschlandradio, and Deutsche Welle published a joint set of principles for the use of AI in editorial processes and reporting. The use of AI is not ruled out, but it is linked to journalistic added value, sustainability, and transparency.
The use of AI is more widespread in private radio, as research by heise shows. Some private radio stations play mainly AI-generated songs, especially at night. This is primarily due to commercial reasons, as no royalties are payable to the Society for Musical Performance and Mechanical Reproduction Rights (GEMA) for AI-generated music.
(mho)
Leave a comment