Imagine a digital lie detector that works for text, images, and videos all at once. It compares suspicious posts against a massive map of known facts to see if they match. Instead of just guessing, it gives a 'fakeness score' to help people spot manipulated content in real-time.
Digital platforms struggle to identify manipulated multimodal content (video/image/text) in real-time. This leads to the rapid spread of disinformation, eroding user trust and democratic stability.
An open-source debunking API and four interfaces: a web plug-in, a collaborative platform (Disinfopedia), a smartphone app, and an AR/VR interface.
If you are a platform operator dealing with viral fake news — this project developed a debunking API that provides a 'Disinfoscore' to flag manipulated content. This allows for faster moderation of multimodal signals including video and images.
If you are an EdTech provider dealing with low digital literacy in students — this project developed playful, human-centered interfaces for educational purposes. These tools help students become discerning consumers of information through interactive debunking.
If you are a security firm dealing with state-sponsored propaganda — this project developed multimodal knowledge graphs to map disinformation sources and diffusion patterns. This enables the detection of coordinated attacks like those seen in the war in Ukraine.
Based on available project data, no pricing or cost information is provided as the project focuses on developing an open-source debunking API.
The project is an Innovation Action involving 14 partners and 4 industry entities, suggesting a design intended for real-world application across web plug-ins, apps, and AR/VR interfaces.
The project explicitly mentions the development of a 'first-of-its-kind open-source debunking API', implying an open-access licensing model.
Integration is achieved through a web plug-in for browsers and social media, a smartphone app, and a collaborative platform called Disinfopedia.
The project period runs from 2024-01-01 to 2027-12-31, indicating that full deployment and validation occur within this four-year window.
The consortium is well-balanced for a technology transfer project, consisting of 14 partners across 8 countries. With a 29% industry ratio (4 companies, including 2 SMEs), there is a strong bridge between the 6 universities and 3 research centers. This mix ensures that the academic AI research (NLP, RNN, CNN) is grounded in practical software development and media professional requirements.
Contact the University of Latvia (LATVIJAS UNIVERSITATE) as the project coordinator.
Contact SciTransfer to explore licensing opportunities for the open-source debunking API.
