4a Data governance, analytics and processing

The emergence of VW as persistent, intelligent, and interconnected digital ecosystems necessitates a radical rethinking of how data is governed, processed, and analysed. Unlike traditional internet platforms, VW are immersive, dynamic, and designed for multiple users, incorporating real-time AI driven interactions, biometric feedback, decentralised content generation, and cross-domain services.

In such environments, data governance should be considered as a foundational design principle that ensures trust, inclusivity, security, compliance, and ethical alignment. This necessitates robust frameworks for interoperability, access control, consent management, and auditable data flows, consistent with European regulatory frameworks such as the General Data Protection Regulation (GDPR), the AI Act, the Data Act, the Data Governance Act, and forthcoming Digital Identity regulations.

Equally important are advanced data analytics and novel data architectures to support capabilities such as context-aware behaviour modelling, content personalisation, moderation, and economic system simulation.

Finally, the strict demands of VW, particularly in terms of real-time responsiveness, adaptiveness and scale-up, require seamless integration of data processing pipelines with diverse computing infrastructures, including edge, cloud and HPC. These systems must operate in real-time and across geographies, while maintaining end-to-end low latency, data security, and energy efficiency.

Key research topics for data governance, analytics and processing

To build future-ready and ethically grounded VW, several research themes must be addressed:

4a.1 Data formats and interoperability

• End-to-end data interoperability across VW components, ensuring technical interoperability in all VW elements, and semantic interoperability in those intended to interpret and use the data
• Seamless deployment of data pipelines over heterogeneous infrastructures, including adaptive orchestration across edge, cloud, fog, and HPC environments, ensuring low latency, resilience, and energy efficiency.

4a.2 Data governance and compliance

• Point-to-point data governance, including traceability, privacy, data quality, consent management, identity assurance, and security
• Ethical and legal compliance by design, embedding compliance, fairness, explainability, and inclusiveness into data practices from the design phase, aligned with evolving European regulations.

4a.3 Data processing

• Secure, scalable, and federated data processing architectures, specifically conceived to address VW requirements, including privacy preserving computation (federated learning) and architectures to handle sensitive, distributed, and high-volume data securely.
• Domain-specific data preparation and representation, including techniques to pre-process, clean, annotate, enrich, … data for its specific use in VW related applications, such as DTs, AI agents, content generation, behavioural simulation, …
• Integration and analysis of multimodal and real-time data, to support context-aware AI, personalisation, and adaptive experiences.

Relevant EU programs and initiatives contributing to the realisation

The European Data strategy, published in 2020 paved the way to fully unlock the value of data in Europe, through the European Common Data Spaces. The implementation of these data spaces, funding under the Digital Europe Programme, initially targeted eight relevant domains in the EU, including media (Trusted European data Space for Media / TEMS). Additionally, the Horizon Europe Programme is funding research and innovation actions to advance the state of the art in topics like compliance, privacy preservation, green and responsible data operations, data management, data trading, monetizing, exchange and interoperability, architectures and standards for complex data cycles, or AI-driven data operations and compliance technologies. This strategy will be reinforced with the upcoming European Data Union strategy, that will rely on three pillars: facilitate or increase the access to data (specially for AI), regulatory simplification and international data sharing

In July 2024, the European Commission’s AI Innovation Package introduced AI Factories, funded by the EuroHPC JU. These open ecosystems, built around EU supercomputers, are designed to develop and train large AI models, offering AI and data services and accessing vast amounts of high-quality industrial data. Complementing this, the AI Continent Action Plan proposes Data Labs to ensure AI developers have access to sector-specific, high-quality data.

Lastly, significant standardisation efforts, like CEN-CENELEC’s JTC25 “Data, dataspaces, cloud and edge”, aim to establish harmonized standards supporting interoperability, trust, and scalability in European data ecosystems, in line with the Data Act.

From the private side, BDVA has been supported the development of the EU data economy, first as the private counterpart of the EC for the BDV PPP, then as part of ADRA for the ADR-PPP. BDVA joined forces with IDSA, FIWARE and Gaia-X to form the Data Spaces Business Alliance. BDVA is a private member of the EuroHPC JU and of the Computing Continuum Initiative.