Periodic Reporting for period 1 - ENFIELD (ENFIELD: European Lighthouse to Manifest Trustworthy and Green AI)
Reporting period: 2023-09-01 to 2025-02-28
ENFIELD operates at the intersection of four scientific pillars: Green AI, Adaptive AI, Human-Centric AI, and Trustworthy AI, reflecting Europe’s commitment to ethical, sustainable, and impactful AI development. These pillars are strategically aligned with application verticals critical to European society and industry: healthcare, energy, manufacturing, and space. ENFIELD’s overall objectives include:
• Driving AI research within four core scientific pillars; tailoring solutions to the specific challenges and needs of the four verticals.
• Developing a Common Research Vision and Roadmap to guide AI innovation in Europe.
• Strengthening AI skills and capabilities across Europe through education, training, hackathons, summer schools, and MOOCs.
• Building and sustaining a vibrant, multidisciplinary, and cross-sectoral AI ecosystem that unites academia, industry, and policy stakeholders.
• Ensuring AI systems' trustworthiness and security through a robust Safety and Security Risk Assessment Framework.
• Providing financial support to third parties via open calls to catalyze innovation and expand the project's impact.
The political and strategic context is defined by the EU’s drive to secure digital sovereignty, foster responsible AI innovation, and implement the Artificial Intelligence Act. ENFIELD aligns closely with the EU’s Green Deal, the Digital Decade targets, and initiatives such as AI-on-Demand and the European AI Lighthouse.
ENFIELD’s impact strategy is multi-layered:
• Scientific impact: Through over 75 novel AI solutions (algorithms, methods, datasets, prototypes), 180 scientific publications, and extensive collaborations, ENFIELD advances the state of the art in AI.
• Societal and economic impact: By fostering ethical AI practices and supporting innovation in critical sectors, the project will directly benefit European citizens and industries. The education and training strategy targets the next generation of AI talent, enhancing employability and research capacity across Europe.
• Policy and regulatory impact: ENFIELD actively informs European AI policy and standardization, notably through its White Paper on Green and Trustworthy AI and engagement with regulatory bodies.
• Sustainability impact: The focus on Green AI reduces AI's environmental footprint, contributing to EU climate goals.
• Driving AI research within four core scientific pillars; tailoring solutions to the specific challenges and needs of the four verticals.
• Developing a Common Research Vision and Roadmap to guide AI innovation in Europe.
• Strengthening AI skills and capabilities across Europe through education, training, hackathons, summer schools, and MOOCs.
• Building and sustaining a vibrant, multidisciplinary, and cross-sectoral AI ecosystem that unites academia, industry, and policy stakeholders.
• Ensuring AI systems' trustworthiness and security through a robust Safety and Security Risk Assessment Framework.
• Providing financial support to third parties via open calls to catalyze innovation and expand the project's impact.
The political and strategic context is defined by the EU’s drive to secure digital sovereignty, foster responsible AI innovation, and implement the Artificial Intelligence Act. ENFIELD aligns closely with the EU’s Green Deal, the Digital Decade targets, and initiatives such as AI-on-Demand and the European AI Lighthouse.
ENFIELD’s impact strategy is multi-layered:
• Scientific impact: Through over 75 novel AI solutions (algorithms, methods, datasets, prototypes), 180 scientific publications, and extensive collaborations, ENFIELD advances the state of the art in AI.
• Societal and economic impact: By fostering ethical AI practices and supporting innovation in critical sectors, the project will directly benefit European citizens and industries. The education and training strategy targets the next generation of AI talent, enhancing employability and research capacity across Europe.
• Policy and regulatory impact: ENFIELD actively informs European AI policy and standardization, notably through its White Paper on Green and Trustworthy AI and engagement with regulatory bodies.
• Sustainability impact: The focus on Green AI reduces AI's environmental footprint, contributing to EU climate goals.
1. WP2 – Research Excellence
• The consortium launched structured research activities across these pillars, with 54 active topics tracked under 13 main research areas by M18.
• An adaptive process of monitoring, merging, and closing research topics ensured responsiveness to evolving challenges.
• Integration with industrial needs progressed through initial alignment of research outputs with the verticals, particularly strong synergies between Green AI and the Energy/Manufacturing sectors, and Adaptive AI with Manufacturing and Space.
• Deliverables D2.1 and D2.2 captured initial and intermediate findings, setting a robust scientific foundation.
2. WP3 – Societal Relevance
• Key deliverables D3.1 and D3.2 mapped out detailed use cases for each vertical, establishing clear links between research and practical implementation.
• Nine third-party TIS projects were launched to address specific technical challenges within each vertical (e.g. Improving low data volume defect detection via neural architecture search).
• Significant progress was achieved in co-creation, with online and in-person workshops enabling deep engagement between research teams and industrial stakeholders, driving technical refinement and practical application.
3. WP4 – Common Vision and Roadmaps
• The Education and Training Strategy (D4.1) laid out a roadmap for upskilling Europe’s AI talent.
• The Common Research Vision and Roadmap (D4.2) and the initial Safety and Security Risk Assessment Framework (D4.3) were completed.
• The DIHIWARE platform was populated with a Catalogue of Learning Resources and updated continuously to support both internal and external stakeholders.
• Practical events included a hackathon on AI bias mitigation and a multidisciplinary webinar.
• The consortium launched structured research activities across these pillars, with 54 active topics tracked under 13 main research areas by M18.
• An adaptive process of monitoring, merging, and closing research topics ensured responsiveness to evolving challenges.
• Integration with industrial needs progressed through initial alignment of research outputs with the verticals, particularly strong synergies between Green AI and the Energy/Manufacturing sectors, and Adaptive AI with Manufacturing and Space.
• Deliverables D2.1 and D2.2 captured initial and intermediate findings, setting a robust scientific foundation.
2. WP3 – Societal Relevance
• Key deliverables D3.1 and D3.2 mapped out detailed use cases for each vertical, establishing clear links between research and practical implementation.
• Nine third-party TIS projects were launched to address specific technical challenges within each vertical (e.g. Improving low data volume defect detection via neural architecture search).
• Significant progress was achieved in co-creation, with online and in-person workshops enabling deep engagement between research teams and industrial stakeholders, driving technical refinement and practical application.
3. WP4 – Common Vision and Roadmaps
• The Education and Training Strategy (D4.1) laid out a roadmap for upskilling Europe’s AI talent.
• The Common Research Vision and Roadmap (D4.2) and the initial Safety and Security Risk Assessment Framework (D4.3) were completed.
• The DIHIWARE platform was populated with a Catalogue of Learning Resources and updated continuously to support both internal and external stakeholders.
• Practical events included a hackathon on AI bias mitigation and a multidisciplinary webinar.
Key results include
• A dynamic network of 54 active research topics aligned to real-world use cases and industrial needs.
• The deployment of 9 third-party TIS projects and the successful onboarding of 36 TES researchers funded through 28 TES projects.
• One patent application filed with the EPO.
• Three datasets generated.
• Foundational frameworks such as the Common Research Vision and Roadmap (CRVR), the Safety and Security Risk Assessment Framework (SSRAF), and initial deliverables on AI education and training strategies.
• Technical guidance and innovation mentoring for third-party projects.
Potential Impacts
1. Scientific Impact: ENFIELD is pushing forward AI research frontiers, addressing challenges in AI safety, energy consumption, adaptability, and ethical alignment. The anticipated delivery of 180+ peer-reviewed publications and 75+ new AI methods and prototypes will cement ENFIELD’s contribution to global AI research.
2. Economic and Industrial Impact: By translating research into viable products (e.g. Safe intelligent agent to optimize ship energy management), ENFIELD is poised to stimulate new business models, job creation, and sectoral innovation, particularly in healthcare, manufacturing, energy, and space.
3. Policy and Regulatory Impact: The draft White Paper on Green and Trustworthy AI and contributions to standardization are directly influencing the EU’s AI governance, with aims to ensure harmonization with the AI Act, GDPR, and other key frameworks.
4. Societal Impact: By prioritizing human-centricity, inclusivity, and sustainability, ENFIELD is supporting the EU’s vision of a trustworthy, ethical AI ecosystem that addresses real societal needs, from improved healthcare services to reduced environmental footprints.
• A dynamic network of 54 active research topics aligned to real-world use cases and industrial needs.
• The deployment of 9 third-party TIS projects and the successful onboarding of 36 TES researchers funded through 28 TES projects.
• One patent application filed with the EPO.
• Three datasets generated.
• Foundational frameworks such as the Common Research Vision and Roadmap (CRVR), the Safety and Security Risk Assessment Framework (SSRAF), and initial deliverables on AI education and training strategies.
• Technical guidance and innovation mentoring for third-party projects.
Potential Impacts
1. Scientific Impact: ENFIELD is pushing forward AI research frontiers, addressing challenges in AI safety, energy consumption, adaptability, and ethical alignment. The anticipated delivery of 180+ peer-reviewed publications and 75+ new AI methods and prototypes will cement ENFIELD’s contribution to global AI research.
2. Economic and Industrial Impact: By translating research into viable products (e.g. Safe intelligent agent to optimize ship energy management), ENFIELD is poised to stimulate new business models, job creation, and sectoral innovation, particularly in healthcare, manufacturing, energy, and space.
3. Policy and Regulatory Impact: The draft White Paper on Green and Trustworthy AI and contributions to standardization are directly influencing the EU’s AI governance, with aims to ensure harmonization with the AI Act, GDPR, and other key frameworks.
4. Societal Impact: By prioritizing human-centricity, inclusivity, and sustainability, ENFIELD is supporting the EU’s vision of a trustworthy, ethical AI ecosystem that addresses real societal needs, from improved healthcare services to reduced environmental footprints.