Gen AI Innovation Challenges Workshop with ADRA
AI-BOOST and ADRA – AI-Data-Robotics-Association co-organised the Gen AI Innovation Challenges Workshop, focused on leveraging Generative AI to boost European industrial competitiveness, emphasising the necessity of aligning technological development with core European values like trust, sustainability, and inclusion. The event featured industry use-cases, insights from the EC, a winner from the Large AI Grand Challenge, Adra and AI-on-Demand Platform presentations and an amazing roundtable full of thoughts from field experts.
Philip Piatkiewicz, the Secretary General of Adra , provided opening remarks focusing on Adra’s vision, mission, goals, and strategic initiatives, particularly in relation to the European “Applied AI Strategy.”
Key takeaways from the European ‘Apply AI’ Strategy
The workshop introduced the new Apply AI Strategy recently launched by the European Commission (EC), which advocates for an “AI first” policy across strategic European sectors. Evangelia Markidou from the EC highlighted that the strategy focuses on sectoral needs to ensure policies and are tailored to specific industrial realities.
It was also said that the Apply AI Strategy sets the scene for the ongoing of the AI-BOOST Gen AI Innovation Challenges, and were called upon industrial representatives of all strategic sectors to come forward with use-cases to help define them, ensuring the solutions developed are relevant and impactful for European industry.
AI-BOOST project and the Gen AI Innovation Challenges
The workshop’s main objective was to promote the AI-BOOST project and its ongoing search for challenge owners for the Gen AI Innovation Challenges, a key initiative to boost the European Artificial Intelligence ecosystem and achieve European technological autonomy.
The challenges will focus on strategic needs in high-impact sectors aligned with the ‘Apply AI’ Strategy, presenting an investment opportunity for industry representatives.
The project is looking for private companies, public organisations or non-profit/research entities to propose real-world, high-impact challenges that can be tackled by European innovators, researchers, and startups. By becoming a challenge owner, industry representatives will:
- Co-Define the Challenge: Help shape the AI challenges, contributing to problem definition, selection criteria, and validation scenarios.
- Access Top Talent: Gain access to top researchers, SMEs, students, and startups.
- Become Pioneers: Position themselves as leaders and innovators in their sectors and in the development of Generative AI solutions.
Industrial use cases and application needs
Panellists from diverse sectors illustrated how GenAI can address complex industrial challenges, often demanding small, fast, and sovereign models:
- Mobility (Margriet Van Schijndel-de Nooij, TUE Eindhoven): GenAI is critical for generating massive amounts of synthetic scenarios to train and validate automated driving functions, drastically reducing the need for real-world driving and focusing on safety-critical ‘edge cases’.
- Critical Raw Materials (Nabil Belbachir, NORCE): To reduce Europe’s vulnerability to supply chain disruptions, GenAI is needed for urban mining – automating the detection, sizing, and sorting of critical raw materials (including rare earth minerals) from waste streams.
- Manufacturing (Mostafizur Rahman, MTC): GenAI is being piloted for applications like text-to-parametric CAD design and is essential for developing an Agentic AI framework to streamline complex factory process chains, especially in managing diverse and legacy data formats.
- Defense & Enterprise (Gorka Unamuno, Multiverse): Focused on making AI “small, fast, and sovereign,” Multiverse uses quantum-inspired mathematics to create ultra-compressed nano-models deployable on the edge (e.g., FPGAs for real-time ship detection in earth observation) and for efficient enterprise tools like LLMs for internal technical manuals.
Roundtable: The European Way Forward - Sovereignty, Strategy, and Scaling AI
Hosted by Valdimir Estivill from UPF, the roundtable gave the speakers the opportunity to reflect in two major questions:
- How can it be ensured that Generative AI flourishes in Europe with European values? and
- What is the AI-BOOST Gen AI Innovation Challenges contribution?
The first question explores the relationship between adhering to European values (Trust, Sustainability, Inclusion) and driving GenAI development. The consensus was that these values must form the foundation of a sovereign European AI strategy. The main hurdle is Europe’s vulnerability due to a lack of control over the global AI supply chain (e.g., hardware/GPUs). To lead, Europe must focus on quality and ethics over model size, mandated by three principles:
- Trust by design: Built-in accountability and human oversight.
- Sustainability by design: Energy efficiency as a core KPI.
- Inclusion by design: Wide access to computing and data for all European actors.
The path forward requires coordinating existing talent and resources to build a unified, sovereign supply chain.
The second question addressed how challenges and competitions as promoted by AI-Boost, contribute to creating synergies and accelerate GenAI/technological development within Europe.
- Mobilising Talent: Connecting top European talent (SMEs, startups) with essential high-performance computing resources (e.g., EuroHPC’s supercomputers).
- Driving Impact: Ensuring research addresses realistic, high-impact industry problems, resulting in useful output rather than purely academic exercises.
- Fostering Ecosystems: Creating necessary cross-sectoral collaboration among researchers, industry, and specialists to enable knowledge transfer and synergy across sectors (Health, Mobility, etc.).
- Validation: Serving as a “test lab” to validate solutions in real-world scenarios, positioning winners as pioneers in Generative AI.
Tilde: The Large AI Grand Challenge Winner
Andrejs Vasiļjevs from Tilde (one of the winners of the AI-BOOST Large AI Grand Challenge) highlighted how the competition was the crucial catalyst for them to build a LLM that tackles the critical problem of global AI models being heavily English-centric.
By awarding them with 2 million GPU hours on the LUMI supercomputer, the Large AI Grand Challenge provided the necessary computing infrastructure to transform Tilde’s linguistic expertise into a concrete, world-class model that addresses Europe’s unique linguistic and ethical requirements.
This enabled the creation of the TildeOpen, a 30-billion-parameter open-source model built for Europe’s languages, now available on Hugging Face.
The role of AI-on-Demand Platform
Almost by the end of the workshop, Heli Harrikari from DIMECC presented the AI-on-Demand (AIOD) Platform, which will serve as a strategic hub for the European AI ecosystem. This platform is designed to:
- Connect Resources: Link EU-funded projects, Digital Innovation Hubs, supercomputers, and industrial resources.
- Offer a Unified Value Chain: Provide a seamless “Build, Train, and Operate” environment with low-code tools and consultancy services.
- Ensure Trust: Be “European by design,” aligning with the EU AI Act and Data Act to provide a secure, vendor-independent marketplace.
Wrapping up & next steps
The consensus of the event was clear: Europe possesses the talent, resources, and policy framework to lead in GenAI, provided it accelerates collaboration and focuses on its unique strengths – building trustworthy, efficient, and socially aligned technology to solve real-world industrial challenges.
As per the Gen AI Innovation Challenges, they are currently ongoing with an open Expression of Interest for Challenge Owners.
To know more consult the AI-BOOST website and follow the project on LinkedIn.
—-
Related Content
---
Follow AI-BOOST today to take part in the shaping of the next level of European AI open competitions.
Join AI-BOOST’s community on LinkedIn (@aiboost-project) & on Youtube (@aiboost-project)