AI in journalism: concerns prevail among German citizens
Deception and a lack of transparency are the focus of criticism of AI in the media. Respondents were particularly skeptical of articles written entirely by AI.
(Image: Daniel AJ Sokolov)
Three quarters of Germans (76%) are concerned about the credibility of the media when artificial intelligence (AI) is involved. They assume that the technology leads to a decline in trust in news and media content. This is a key finding of the "Transparency Check" study on the "Perception of AI journalism", which the state media authorities published on Wednesday. It is based on a representative online survey of 3013 internet users. According to the survey, 56% of participants even see AI as a threat to democracy in Germany. At the same time, over 90 percent consider clear rules for the use of technology in the media and labeling to be essential.
According to the analysis, concerns about AI in journalism predominate across all age groups. Deceptions such as artificially generated but authentic-looking deepfakes and a lack of transparency are the focus of criticism. German citizens are particularly skeptical of purely AI-generated content such as articles written entirely with technology or synthetic moderator voices. Younger, more formally educated users who consider themselves to have a high level of media literacy, on the other hand, are more likely to see opportunities in AI. According to them, such automated tools could help with research or fact-checking.
AI already widespread in the media
A test conducted by the study makers revealed that respondents sometimes find it difficult to recognize AI-supported articles. In the case of texts in particular, 67 percent of respondents still recognize the use of technology based on transparent indications. "Virtually no one, on the other hand, recognizes the speaker's voice in audio contributions," it continues. There is also potential for improvement in the labeling of video contributions. Younger viewers are most likely to respond to the information here. The most important recognition signals in video are – more important than the label itself – an artificially automated-sounding voice and tonality in the contribution as well as the use of avatars.
Videos by heise
In general, the majority (57%) rate their own knowledge of AI as poor. 41 percent have already tried out tools such as ChatGPT or DeepL. "The use of AI in journalism requires more than just technical expertise," summarizes transparency check manager Christian Krebs. "It's about an ethical attitude that focuses on responsibility and transparency." Media companies and platforms must "actively promote clear labeling and comprehensible processes".
According to a survey by the Media Association of the Free Press (MVFP), 78% of German publishers see AI as a megatrend. The technology is already being used in one in three magazine editorial offices for researching, selecting topics, creating and editing texts. The German Journalists' Association (DJV) warns: "Artificial intelligence must not be integrated into editorial processes at a crucial point." The task of journalists is much more than just fine-tuning.
(vbr)
