Europeans: Traditional media, AI detectors, and the state must check AI content

Contents

From the perspective of Europeans, the protection of democracies in Europe from AI fake news and manipulation attempts can be ensured by a triad of checking institutions: traditional media, AI detectors, and state regulation. Personal media literacy is rated as less effective. This is according to a current study by the Vodafone Institute, the think tank of the telecommunications service provider. AI is thus considered a major threat to the manipulation of elections and also to the undermining of trust in political institutions through generated fake news or disinformation.

Importance of traditional media has increased

For the study, 12,086 eligible voters aged 18 and over in twelve European countries were surveyed via computer-assisted online interviews in spring 2025, including 1002 people from Germany. According to the survey, for more than half of Europeans, regardless of age, the importance of editorial reporting has increased due to the rising threat of fake news. A total of 53 percent of respondents agreed with the statement that the importance of editorial contributions has increased due to the current situation. Among 25- to 34-year-olds, it is 57 percent, and among 18- to 24-year-olds and 35- to 44-year-olds, it is slightly more at 54 percent.

Older people consume traditional media anyway, the study shows: on average across Europe and all age groups, around 60 percent of respondents turn to traditional media such as television and radio news for political information. The use of television and radio increases with age. The proportion in the 18- to 24-year-old age group is just over 40 percent. In addition to television and radio news, other sources play an important role: personal exchange within the family, among friends, or at work (46 percent), editorial reporting (43 percent), and political talk shows and documentaries on television (41 percent).

60 percent of 18- to 24-year-olds use social media for political information. Instagram (61 percent), YouTube (57 percent), and TikTok (56 percent) are the main platforms that have been used for this purpose at least once. Looking at all respondents across all age groups, there is a divided picture for online news consumption: 47 percent get their information from news portals, and 42 percent from social media.

Younger people perceive more fake news

The frequency with which people encounter fake news also differs between generations. Accordingly, among 18- to 24-year-olds, 38 percent stated that they had encountered fake news very or extremely frequently in recent months, while the figure for those over 64 was 16 percent. The study speculates that this could be partly because young people spend more time online and therefore have more contact with disinformation. Whether they are also better at recognizing fake news was not investigated by the study. However, it comments elsewhere: "While most are convinced that they are not influenced by fake news, the risks of manipulation are assessed as much greater for other people: 39 percent compared to 18 percent. This is a well-known phenomenon, and therefore not surprising that it is found in all countries. However, it is quite relevant, as it can contribute to social division. If individual groups believe that 'the others' are being manipulated by the media, this can increase mistrust in society."

28 percent of all respondents stated that fake news poses a very large or extreme threat to democracy. People from Spain are most convinced of this with 38 percent agreement, and people from Finland least so with 15 percent. The younger the respondents, the more strongly they assess the influence of false information on democracy (35 percent among 25- to 34-year-olds vs. 21 percent among those over 64). The frequency with which fake news is perceived also differs significantly between countries: in Estonia and Greece (39 and 35 percent), respondents perceived significantly more fake news than, for example, in Finland and Sweden (20 and 19 percent).

AI as an accelerant for fake news

Since AI can also be used to create fake news and deceptively real videos and photos in a political context, participants were also asked about this: The greatest risks of AI for democracy are seen in opportunities for election manipulation in the run-up to elections (36 percent) and the undermining of trust in political institutions through AI-generated fake news (30 percent). Currently, skepticism also prevails among European respondents regarding the perception of AI. Only about 30 percent believe that the advantages of AI outweigh the disadvantages overall. Respondents from Germany and the United Kingdom are particularly critical of AI content, while there is more trust in Southern Europe.

According to the respondents, a triad of measures has the greatest future potential to counter this: AI detectors, traditional media, and state regulations. 45 percent see verification by traditional media as particularly relevant, which was mentioned particularly by older people. Recognition algorithms in social media (35 percent), media-literate users (31 percent), and the promotion of best practice exchange by international organizations (23 percent) are considered less relevant for the future.

Accordingly, respondents advocate for measures such as labeling AI-generated content (65 percent), fact-checking (59 percent), and the use of national supervisory authorities (58 percent). Participants from Portugal are strongest in advocating for the various aspects of the DSA and the AI Act in this context with 70 percent, while people from Poland do so least with 50 percent. Furthermore, support for regulatory measures increases with age and formal education. For example, fact-checking and the suppression of hate speech are supported by almost half of those with low formal education (47 percent), while two-thirds of those with high formal education (65 percent) support them.

Furthermore, 43 percent of all respondents consider tools for detecting artificially created or altered videos, images, or audio files (AI detectors) to be promising solutions. So far, only 14 percent of respondents use such detectors, but interest is greater among young and formally highly educated people. While the willingness to use them varies between countries, almost a third of Europeans can well imagine using them to verify political content. Overall, the study shows: Older respondents rely more on state regulations and verification with traditional media (49 percent and 56 percent of those over 64, respectively), while younger people consider AI detectors to be more suitable (50 percent of 18- to 24-year-olds).

AI more of a topic for younger people

According to the study, younger people are the ones who engage more with AI tools and have already used them for political topics. The younger the respondents, the greater the willingness for (at least occasional) use. Among 18- to 24-year-olds, it is 79 percent, while among those over 64, it is only 24 percent.

By far the leader among AI tools is ChatGPT, which has already been used by a third in a political context at least once, followed by Google Gemini, Meta AI, Copilot, and DeepSeek. However, these figures should not be overestimated, because as the study also clarifies, the use of AI tools for political opinion formation lags far behind traditional media or news portals and social media: only 11 percent of respondents regularly use ChatGPT or other AI tools when it comes to obtaining political information.

And even though younger respondents aged 18 to 24 believe that AI brings more advantages than disadvantages by 48 percent, these age groups question AI-generated content slightly above average at 43 percent. Across all age groups, 40 percent of respondents are skeptical of AI content, similar to social media content at 41 percent. With 43 percent, only party advertising fares worse among respondents.

(kbe)