Generative AI puts trust in the news media to the test

The spread of fake news and disinformation has steadily risen in past years, with audiences increasingly getting their news online, where false information – sometimes generated by AI – spreads faster, threatening audiences’ trust in media.

In this series of infographics, Euractiv takes a deeper dive into the trends playing out across Europe’s media landscape.

AI and mistrust in media

The trend of consumers trusting the media less and less has been observed for a while, resulting in some turning to social media to get their news instead. Both this year and last year, in the US, trust in media hit all-time lows. This pattern is the same in the EU, as laid out in the graph below. Meanwhile, TikTok, for example, is popular among the younger generation as an alternative to more traditional news sources.

Using data from 2020, the above chart also shows that social media plays a big role as a news source.

Yet, according to a Eurobarometer survey published by the European Parliament on 17 November which uses 2022 data, the main source of news is television, with 71% of the respondents saying that it was one of their most used sources during the past seven days at the time of the research. However, there is an 11% increase in those using social media platforms to get news compared to last year’s survey.

To counteract the risks of artificial intelligence, some believe AI can help debunk fake news, such as with the Horizon Europe project and AI4Trust. AI4Trust is developing a system “based on machine-human cooperation and advanced solutions”, while Horizon Europe’s 2023/24 Work Programme aims “to move closer to an AI-based market” and solutions to ensure “access to meaningful information, quality content and trustworthy online interaction”.

However, on social media platforms, it can also be the source of it. Not knowing what is real and what is AI-generated does not help trust the news.

In March, for example, the EU law enforcement agency Europol warned that while ChatGPT can benefit businesses and individuals, malicious actors can exploit such technologies, posing a law enforcement challenge.

“Criminals are typically quick to exploit new technologies and were fast seen coming up with concrete criminal exploitations, providing first practical examples mere weeks after the public release of ChatGPT,” reads the report.

“The dissemination of nude images of minors created with AI has been a concern for law enforcement for a while, and one bound to become increasingly more difficult to address,” Europol also told Euractiv in October.

The law enforcement agency added that AI-generated material must be detected using AI classifier tools, but currently, they are not authorised to use classifier tools for this specific purpose.

Online news sources

In the European Union, those reading news online increased in the past years, as the graph below shows, between 2013 and 2022, according to Statista.

While online, the frequency and the types of fake news readers can differ, as seen in the graph below, which covers Europe’s largest countries.

Not only does the frequency of seeing disinformation depend on the different types of media, but citizens’ exposure level to disinformation, in general, is not the same in every country either. While each country may have a different percentage regarding fake news, the following map shows that Scandinavian countries are less affected than those in central and southern Europe.

Last year, Bulgarian citizens had the highest exposure to misinformation at 28% between April 26 and May 11, while those living in the Netherlands were least exposed to it with 6%.

Regardless of frequency or exposure, seeing false information can impact one’s life differently. Most claim (35%) to check the news more thoroughly after such an encounter. However, almost as many people (33%) believe it does not impact them.

[Edited by Alice Taylor/Nathalie Weatherald]