EUROPE / ARTIFICIAL INTELLIGENCE

AI in Central European Newsrooms: Thomson's New Study Reveals Insights

AI’s integration into journalism has sparked a mix of enthusiasm and scepticism across Central Europe, according to research led by Thomson Foundation in collaboration with the Media and Journalism Research Center. On one hand, AI Tools promise to streamline operations, enhance content personalisation and improve data management. On the other hand, they raise ethical concerns, especially regarding the creation and dissemination of disinformation.

The research carried out during March and April 2024 indicated the average percentage of AI usage in the media across Central Europe does not exceed 15%. Its use mainly involves technical tasks, while more complex operations are still performed by journalists and editors. This is the case especially in small newsrooms that still lack enough staff and technical capacity to use AI tools in their work more actively. The formation of AI teams in most newsrooms is a new development. The number of media across the region employing AI experts is still very small which corresponds with the limited knowledge of AI technologies in their newsrooms.

This was especially indicative for Thomson’s team in the region who work strategically to help smaller outlets develop their business and production capacities.

“Our main intention was to identify the trends of the AI use by media organizations globally and specifically in the region of our interest – Central Europe,” says Davor Marko, Western Balkans and Central Eastern manager for Thomson. “We were also assessing media experience, but also capacities and potential to work more in this area and use the knowledge and evidence-based recommendations in order to tailor our support programs in the future.”

 AI across the region

The comparison across the region reveals a nuanced picture of AI’s integration into journalism.

In the Czech Republic, journalists worry that AI could boost disinformation, potentially distract the public from key issues and manipulate voting behaviour. Similarly in Hungary, social media platforms have become battlegrounds flooded with propaganda, further exacerbated by the proliferation of fake profiles.

In Poland, fact checkers and media watchdogs rely on AI tools for tasks like comment moderation and content analysis to tackle disinformation.

In Slovakia, some media outlets use AI to generate content based on real-time results. In the recent elections, AI helped generate hundreds, sometimes thousands, of articles about the results in each Slovak municipality. There’s also recognition that generative AI, particularly in the form of deep fakes, could negatively impact public trust in elections.

 The adoption of AI in Central European newsrooms presents both opportunities and challenges, as media organisations indicated. While AI can increase efficiency in news organisations, media teams fear the potential effect of AI on professional journalism.

AI in media across the CE region is often used for: personalised content delivery to users; speech-to-text conversion; video content creation; social media analysis; text editing and moderation. The most widespread used tools are Open AI’s ChatGPT, Microsoft’s Copilot and Midjourney.

Despite the existence of AI tools for content creation, these are not widely used by media in Central Europe, or their use is in an experimental phase. The main problem of large language models is factual unreliability and the production of different answers to the same tasks over time. This makes them unsuitable for producing unsupervised content.

The other limitation concerning the more widespread use of AI is the financial cost, especially for more advanced tasks, such as the creation of the audio version of an article.

Several of the media surveyed wanted to maintain the “human touch” in supervision and moderation of content.

National perspectives, media independence, and political contexts play a role in how AI is integrated in newsrooms. Czech journalists are overall cautious, stressing the potential for AI to perpetuate disinformation especially in vulnerable political contexts. In Hungary, current AI usage for disinformation remains largely undocumented. Cheap fake videos dominate political campaigns in the country but in the country’s media the use of AI tools specifically for content creation are not common. However, among the existing examples Hirstart, a news aggregator, ventured into AI-produced podcasting as early as May 2020.

In contrast, Polish media are progressively embracing AI to enhance journalistic practices and combat disinformation, maintaining a focus on ethical implications. The Slovak media industry is beginning to show interest in AI, concentrating mainly on operational efficiencies such as translation and data processing.

This comparative overview illustrates a diverse spectrum of AI adoption within the region. These trends reflect the need for a balance between technological advancements and ethical journalism practices. 

 The path forward

There is a clear awareness among the respondents in our research that AI technology is the future of media and that it will be necessary to develop human and technological capacities to use it. This transition offers opportunities for collaboration among newsrooms in the region, especially if they aim to standardise AI practices that align with global journalistic standards.

To counter negative developments the report finds, it is necessary to offer training to journalists and media professionals on how to recognise ethical trusted content from AI-generated disinformation. By fostering a culture of continuous learning and adaptation in the use of AI newsrooms will enhance resilience against disinformation and safeguard the integrity of their reporting and upholding journalistic standards in the digital age.

If you want to know more about the benefits of AI for data management, especially for small and independent media outlets, check out our research report in detail here.

 

More on AI

Explore our projects across the globe