Elliot

July 01, 2025 - June 30, 2028

ELLIOT is a European research project that develops the next generation of artificial intelligence (AI) systems, called multimodal foundation models. These are advanced AI models that can understand and connect information from different sources at once, such as text, images, video, audio, and data streams. The project’s goal is to make these powerful AI tools more reliable, transparent, and useful in real-world settings while ensuring they follow European values of ethics, trust, and fairness.

The project focuses on several fields where AI can make a big difference: media and journalism, climate and earth modelling, robotics, mobility (self-driving cars), computer engineering, and workflow automation. By testing AI in these areas, ELLIOT aims to create tools that help people work more efficiently, fight misinformation, and boost creativity, while keeping human needs and safety at the centre.

The role of VRT is to explore how these AI models can support the future of media. VRT will lead experiments in three use cases.

1. Text to 3D: The first one focusses on virtual media production, where we investigate how generative AI can be used for 3D content production.

2. Real-time factchecking: A second use case envolves real-time fact-checking, where we seek to develop a tool that allows editorial teams to fact-check claims being made by studio guests in live television programmes.

3. News avatar: The third use case combines aspects of the other two: virtual production and real-time fact-checking. Here, we explore how a news avatar can operate in a virtual environment.

In addition, the project aims to train these models in a safe way so that data cannot be misused. For this, ELLIOT works with the Flemish branch of OpenChip, which is developing a secure and sustainable AI chip—a European alternative to the American chips made by Nvidia.

The project unites 30 organizations from 12 countries, and is funded by the European Union’s Horizon Europe research and innovation programme under the “Human-Centred and Trustworthy AI” call (HORIZON-CL4-2024-HUMAN-03). This means it is part of Europe’s effort to develop AI that is open, transparent, and aligned with European values.

Relevant link:

• (In Dutch) VRT NWS (2025, 19 juni). VRT werkt mee aan Europees AI-project om toepassingen in de media te verkennen. vrt.be