Polish AI program for radio discontinued after just one week
Off Radio Krakow no longer broadcasts reports and presentations created by AIs. The station believes it has achieved its goal. Ex-employees see that differently
The AI persona "Kuba" is no longer on air at Off Radio Krakow. His image also comes from an image generator.
(Image: Off Radio Krakau)
Instead of three months, the experiment only lasted one week: since last week, the niche station Off Radio Krakow has only been broadcasting music. Previously, personas created by AIs were on air for a week, including voice AI for moderation and fictional interviews with a deceased celebrity.
Editor-in-chief Marcin Pulit said in a statement that the aim of the experiment had been achieved: a debate about AI in radio, public media and journalism in general had been initiated. He is surprised by the sometimes violent reactions. At the same time, Pulit is calling for legal regulation of AI content – There was no talk of this when the experiment was announced.
Moderators rationalized away by AI?
As reported by Deutschlandfunk Kultur prior to the statement, there were already protests on several levels at the beginning of the AI program. A petition by former employees of the station was supported by 20,000 people. It was initiated by journalist Mateusz Demski, who had formerly worked for Off Radio as a music presenter. His job had been replaced by the AI personas. At the beginning of the experiment, the station had already said that only the presenters' temporary contracts had not been extended. The theory of cost-cutting measures through AI moderation is supported by the fact that Off Radio has only been broadcasting music without moderation or spoken contributions since the end of the experiment.
Made-up AI interviews do not reflect the nature of a person
Not only the form, but also the content of the AI program was controversial. This was particularly true of a fictional interview with Polish Nobel Prize winner Wislawa Szymborska. She had died in 2012. On the one hand, the AI made statements about events that the writer could never have experienced.
Videos by heise
On the other hand, the author Remigiusz Grzela told Deutschlandfunk radio that the fictional conversation "does not reflect Szymborska's intellectual greatness, irony and ambiguity". A problem that had already occurred with other AI fakes of people: the models reproduce tone of voice, formulations and other characteristics well, but often not the subtleties such as the irony of a person.
(nie)
