AI4TRUST at the Forefront of AI-Powered Media Verification: Insights from Our Latest Presentation

AI4TRUST at the Forefront of AI-Powered Media Verification: Insights from Our Latest Presentation

Last week, on 13 March 2025, AI4TRUST had the opportunity to present its cutting-edge explainable AI technologies for media analysis and the verification of visual and audio content. The presentation took place at the Data Technology Seminar 2025, hosted by the European Broadcasting Union (EBU), a key industry event, which showcased our advancements to an engaged audience of media and technology professionals.

A Gathering of Media and AI Experts

The event brought together approximately 100-120 participants, primarily from major broadcasting companies such as Deutsche Welle (DW), BBC, the European Broadcasting Union (EBU), ARD, RAI, Netflix, and more. These industry leaders shared insights into their use of AI technologies—ranging from Large Language Models (LLMs) to speech and video processing techniques—for content analysis, recommendation systems, and archiving. Notably, one of our our sister projects, Vera.ai, was involved in the organisation of the event, further highlighting the synergies across European AI initiatives.

Showcasing AI4TRUST’s Technologies

At the initiative of CERTH, Horia Cucu (University Politehnica of Bucharest) did a presentation on AI4TRUST’s explainable AI solutions, which was well received, sparking engaging discussions on combating disinformation and verifying digital content. The audience showed keen interest in several aspects of our work, including:

Platform Adoption: Several participants expressed interest in using our platform. While the details of post-project access are yet to be determined, we anticipate potential integration into upcoming pilot phases.
Handling AI Hallucinations: Technical experts inquired about how AI4TRUST addresses hallucinations in Large Language Models (LLMs) and Vision-Language Models (VLMs), particularly in video anomaly detection. This remains a crucial challenge that we are actively working on.
Engaging Demonstrations: The demonstration videos showcased practical applications for media verification.
Challenges for News Outlets: A thought-provoking question was raised about how AI4TRUST could address challenges faced by news outlets that lose access to information, referencing the case of the Associated Press (AP) being restricted from White House coverage. This is an important consideration that could align with DWS’s disinformation index efforts.

Looking Ahead

The overwhelmingly positive response to AI4TRUST’s presentation reaffirms the critical role of AI in combating disinformation and ensuring content authenticity. As we move forward, we remain committed to refining our technologies, expanding collaborations, and exploring pathways for broader platform adoption.