GenAI4EU: Creating European Champions in Generative AI

Background and scope

While Artificial Intelligence (AI) technologies are already delivering significant improvements in healthcare, aiding climate change mitigation and adaptation, enhancing production systems through predictive maintenance, and increasing the security of European citizens and business, generative AI (GenAI) is expected to revolutionize these fields further.

GenAI promises a transformative impact on our daily lives by significantly advancing capabilities across various domains. GenAI however still faces notable challenges, such as unexplained inaccuracies (i.e., hallucinations) and a lack of transparency and accountability in decision-making, which undermines trust and poses questions regarding the safety and reliability of the systems. These are even more critical in the case of more advanced applications, needing greater scale and integration within user workflows, particularly in sensitive or safety-critical settings like clinical workflows or the management of critical infrastructures. 

In line with the AI innovation package supporting Artificial Intelligence startups and SMEs, and with the EU AI Act, targeted at the development, deployment and take-up of trustworthy AI in the EU, this Challenge contributes to fostering a European, human-centric approach to AI, that embodies EU values and works towards guaranteeing Europe’s tech sovereignty in a critical technology area identified under the Strategic Technologies for Europe Platform (STEP).

The aim of this Challenge is to support start-ups and SMEs committed to bringing transformative AI-driven solutions to market. The proposed solutions should safeguard human autonomy and enhance human expertise thus providing significant added value to decision-making processes, services or industrial workflows. Novel solutions with broad future applicability are particularly encouraged.

Specific objectives

This Challenge will support start-ups and SMEs that are either:

• Further developing and validating new GenAI models; 
• Adapting existing models to specific sectors or types of data where smaller faster and more energy-efficient models would also be applicable; or 
• Integrating and testing the GenAI solutions in existing workflows, and testing these in regulatory sandboxes and real-life settings, including certification and post-market surveillance, as appropriate. 

The developed models must go beyond the state of the art and must look to overcome the current difficulties that limit the extent to which they support human expertise with a particular focus on one of the following areas, and in line with the sectors and applications listed in the AI communication:

• Healthcare: radiology is a frontrunner in the use of GenAI, but current solutions suffer from a lack of trust and integration with clinical workflows, which must be overcome to advance automatic radiology reporting and enhance human-AI interaction while reducing interpretation errors and associated variability. 
• Energy: the integration of GenAI in managing the power grid, alongside the storage and use of renewables by grid operators, for example, calls for high quality, reliable AI systems that deliver safe, real-time decision making to enhance resilience and planning. 
• Security: security professionals including those operating critical infrastructures can scale their work in threat and vulnerability detection and the subsequent response to such threats, with the support of fault-tolerant, high quality and secure AI systems. 
• Public sector: leveraging GenAI to improve the quality, ease of access and efficiency of public services, or to increase the efficiency of public administration services and productivity.
• Cultural and Creative Sectors and Industries: Gen AI systems targeted to the needs of these sectors, catering for cultural and linguistic diversity and ensuring transparency and full respect and recognition of artists and creators rights. 
• Manufacturing: systems to enable mass customisation, enhance sustainability and automation, propose options for design and identify maintenance needs.
• Education: Use of GenAI systems and assistants to enable personalised and adaptive learning experiences, enhancing educational outcomes and accessibility.
• Science: Leveraging GenAI to drive new levels of productivity and capability for researchers in both the private and public sectors, fostering innovation and scientific advancements.

Expected outcomes and impacts

This Challenge will support the further development and validation of Gen AI models and/or downstream applications, which are “European-Value driven” and contributes to the ambitions of the AI Act and the European approach to Artificial Intelligence. The AI models developed under this Challenge are expected to comply with the EU concept for Trustworthy AI and the relevant ethical principles with due attention paid to data quality, transparency & accountability, privacy, and security. In the medium to longer term, it is expected to reduce dependencies and support companies in leveraging the advances in generative AI, enhance their products and develop new ones that:

• Optimise existing workflows: utilising Generative AI technologies to streamline and improve operational processes across various industries;
• Enhance human capabilities: Applying advanced AI models to enhance decisionmaking, creativity, and productivity; 
• Are validated for application: Ensuring that the AI models are both advanced and thoroughly validated for practical applicability and scalability in real-world scenarios.

The selected beneficiaries will receive favourable access to European supercomputing resources for the training of their large foundation models, in the context of the AI Factories in line with the EuroHPC regulation. In addition, they may benefit from additional actions aimed at creating strategic partnerships with major industries or attracting further capital. Where relevant, opportunities may be explored to provide the selected beneficiaries with access to scientific datasets through the European Open Science Cloud or to provide users of the European Open Science Cloud with access to the tools developed by the beneficiaries.

Watch the the info day recording

EIC Work Programme 2025

The EIC Challenges are extensively described in the EIC Work Programme. For more information about this challenge, please go to the corresponding section of the EIC Work Programme.