Poles demand clear rules. 96 percent wants AI material tagged in the media

Over 92 percent Poles are aware of the use of artificial intelligence by journalists, but as many as 66 percent is afraid of manipulation using content created by algorithms and their impact on public opinion – according to a study commissioned by the National Broadcasting Council “Artificial Intelligence in the Media”.

The most important conclusions from the study were presented by the chairwoman of the National Broadcasting Council, Dr. Agnieszka Glapiak, during the conference “AI – media in the era of algorithms”.

“Artificial intelligence is no longer a futuristic promise, but has become a tool that increasingly supports journalists, publishers and digital creators in their everyday work. On the one hand, AI raises curiosity and hopes, on the other hand it raises fears and doubts. To find out Poles' opinions on this subject, the National Broadcasting Council commissioned a study on the perception of AI in the media by the National Research Group. The results are an important point of reference for further discussion on the role of technology and its impact on the quality of the message,” it was stated in the sent. PAP press release.

Respondents want AI marking of materials

As noted, the vast majority of respondents (almost 96%) expect materials created using AI to be clearly marked.

“Despite the cautious approach to artificial intelligence, recipients notice numerous advantages of AI. The most frequently mentioned are: automatic translation of content into another language, saving time in creating and obtaining information and improving the visual side (e.g. graphic materials). More than half of respondents (51 percent of respondents) believe that the use of AI as a tool supporting the process of translating foreign texts increases access to knowledge and information” – emphasized.

“AI Hallucinations”

The strongest concerns, as noted, concern the reliability of the message. Respondents pay attention to the risk of manipulating public opinion and an increase in the threat of disinformation.

“Doubts are also raised about the so-called 'AI hallucinations,' i.e. situations in which algorithms generate erroneous information as facts. Over 44 percent consider such situations to be a threat,” the release emphasized.

“The results of the study commissioned by the National Broadcasting Council clearly show that recipients are not ready to replace presenters in the mass media (television, radio and the Internet) with artificial intelligence. Over 95 percent of respondents are against the use of AI in this role. 76 percent of respondents definitely do not accept presenters generated by AI on television, and over 73 percent said 'no' in relation to radio,” it said.

However, as noted, there is extreme opposition to the use of images of deceased persons in journalistic broadcasts. Over 93 percent respondents reject such solutions, only 2.6 percent expresses acceptance, and 4 percent has no opinion on this subject. Partial acceptance was obtained only in the context of educational and artistic content – 35% approved it. and 25 percent study participants.

Conclusions and recommendations

“The research results allow us to formulate the following conclusions and recommendations regarding the use of AI in the media: clear marking of content created by AI – nearly 96% expect it. respondents (the use of uniform messages and labels is recommended) and emphasizing the role of humans as the final verifier of quality (almost 60 percent of respondents prefer this). And also gradual implementation of AI in supporting functions such as translations and data visualizationwithout replacing creativity (51% of respondents see benefits in this respect) and educating recipients on how to recognize false content – especially in the context of fear of manipulation and disinformation (almost 67 percent of respondents are afraid of the possibility of AI manipulating public opinion),” the release added.

The study, commissioned by the National Broadcasting Council, was conducted by the National Research Group (OGB) on October 23-27, 2025. Method? CATI. Tested representative? sample of N=1000 adult residents of Poland. The sample was selected in a quota-random manner, based on the following criteria: gender, age, education and place of residence (data according to the Central Statistical Office). The analysis covered issues of awareness and attitudes towards AI, benefits and risks, preferences regarding content and presenters, the use of the image of deceased people and the level of trust in technology. (PAP)

akn/ teddy bear/