EU AI Media Platform – modular multilingual public-interest infrastructure for European news and audiovisual media (IA) - Apply AI

Expected outcome

The project will support the development of a modular, trustworthy and interoperable AI-enabled infrastructure enabling pan-European public-interest media platform(s) to make available real time news and information from professional media outlets across the EU and associated countries to wider audiences. It will harness AI to translate, process and distribute content for relevant channels through classification, recognition, linguistic analysis, translation and other multimodal workflows. It will also enable prototypic content creation and testing of immersive and innovation-inspired media experiences consistent with a public-interest mission for different age groups, taking into account gender-responsive design and  promoting news diversity. Overall, it will improve multilingual access, including low-resource European languages, strengthen media pluralism and resilience for public service media and professional media outlets, content creators and independent publishers, and embed trust, provenance and accountability by design.

The concrete expected outcomes are:

• Reliable, modular and interoperable infrastructure with validated service levels and sustainability pathways, integrating AI-based components that support multilingual and multimodal media workflows including audiovisual processing and adaptation workflows, relevant metadata interoperability for broadcast and streaming contexts, social-media-facing interaction and adaptive, personalised, and interactive content experiences (modular journalism) catering for diverse user needs.
• Content in multiple languages and onboarding conditions and equal access for all types of media organisations, from public broadcasters to small local publishers, SMEs, cultural institutions and civil society groups.
• Built-in safety, transparency and governance features, fair human-data interactions (e.g., legibility, agency, negotiability), traceability of key decisions, open evaluation methods, and validated governance and interoperability frameworks.
• Reliable tracking of content origins through provenance information, attribution metadata and verification signals.
  

Scope

This topic will support pilots for AI-enabled media infrastructures, fostering research and innovation in the European media sector. It should deliver innovative solutions by consolidating, integrating and validating reusable AI-enabled infrastructure components and governance arrangements of platform pilots for pan-European public-interest media platform(s), making use, where appropriate, of European AI infrastructure and building on results from relevant Horizon Europe, Digital Europe and pilot actions, in preparation for wider scale-up from 2028 onward. The action will encourage the development, testing and improvement of innovative solutions that respond to the needs of the media ecosystem. Proposals should integrate and scale up relevant results and move existing pilots towards operational validation in real-word conditions against declared service levels, including appropriate resilience, lifecycle-managed storage, a KPI framework and independent evaluation, in line with Union standards, and observability, traceability and content provenance by design. AI components should follow transparent and accountable practices, including documentation, explainability, logging and evaluation, to enable audit, redress and effective oversight.

The project should demonstrate a credible pathway towards broad European language coverage, accessibility and support for relevant emerging formats, with a credible post-grant sustainability approach and, where relevant, onboarding, training and pilot cohorts. It should foster participation of media organisations by addressing industry needs, including bridging tools and common formats linking editorial workflows and platform technologies (e.g. ingestion, aggregation, curation, metadata exchange, rights and licensing parameters, attribution and provenance information). The project should present a mechanism to bring together public and private media outlets and publishers, broadcasters, AI and language-technology providers, platform developers, content creators and social media actors, fact-checking and integrity actors, and rights-management organisations. Where relevant, cultural and heritage institutions and civil society organisations should also be involved. The infrastructure should test advanced open APIs solutions and demonstrate their technical viability. Moreover, the infrastructure should provide SDKs and documentation to enable integration with multiple front ends and third-party environments, onboarding pathways for smaller players, and where relevant, open-source development approaches. This includes social-media-facing interaction and referral functions, and, AI companion-style interactions via an API. Collective contextualisation services, facilitation of user onboarding and support for cooperation, the embedding of robust provenance safeguards for synthetic or manipulated content with clear redress paths and where relevant, cooperation with fact-checking and integrity actors, may equally be considered. The project should contribute to an existing or establish a structured multi-stakeholder governance framework including structured exchanges with relevant stakeholders and related initiatives through an appropriate coordination mechanism.