The future of journalism: SOLARIS hosts a simulation in Rome on mitigating the harm of deepfakes on media content
On May 8-9, 2025, the SOLARIS Consortium held a two-day discussion in Rome, Italy, focused on analyzing and mitigating the negative effects of deepfake content on the work of media and journalists.
The workshop placed participants in a high-stakes scenario involving the viral spread of a deepfake video generated via GANs (Generative Adversarial Networks).
Led by ANSA, one of Europe's leading news agencies, the event recreated a newsroom environment where journalists had to analyze simulated content based on trusted sources, such as government bodies, international organizations, or independent content creators, mimicking real-world scenarios of source evaluation and information verification.
The exercise aimed to test the resilience of current news verification processes under crisis conditions and generate actionable insights at both national and EU levels.
Highlights from the event included operational deepfake analysis sessions, institutional procedure assessments, and a series of "what-if" scenario discussions.
The participants saw projections of fake content and then held a brainstorm session. The covered topics included the effects of deepfake content with plausible information, the institutional procedures around fact checking and the journalists' personal experience with them.
Day two introduced technical demonstrations, such as UNEXE's real-time algorithm for detecting disinformation through sentiment analysis and anomaly detection, and AI-CODE's interactive simulator for safe use of generative AI in journalism.
The event challenged international media, including Economic.bg, AFP, and other experts in examining the editorial chain under emergency conditions and in creating ways to avoid the spread of AI-generated misinformation, thus contributing to a more transparent and health-aware approach to disinformation resilience in the digital age.
The meeting concluded with a collaborative debrief and a draft recommendation document focused on enhancing institutional responses to disinformation and leveraging AI tools without compromising democratic integrity. Participants included representatives from the international press, EU institutions, AGCOM, and research consortia such as IDMO and AI-CODE.
This project is funded by the European Union under Grant agreement No. 101094665. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union. The European Union cannot be held responsible for them.