Strategic Autonomy & Security
- The grant aims to safeguard the Union's strategic assets, interests, autonomy, and security, avoiding technological dependency on non-EU sources.
- For entities established in eligible countries but directly or indirectly controlled by a non-eligible country or entity, specific guarantees are required. These guarantees must demonstrate that participation will not negatively impact the Union's strategic interests, ensure control over the applicant legal entity is not restrained, prevent access by non-eligible countries to sensitive information, and ensure IP ownership remains within the recipient and is not exported without approval.
- Employees or other persons involved in the action may require a national security clearance issued by an eligible country.
- Entities assessed as high-risk suppliers of mobile network communication equipment (or those controlled by them) are explicitly excluded from submitting guarantees and participation.
Ethical & Social Standards
- Proposals must emphasize the development of AI models with trustworthy and ethical behavior and operation, ensuring reliability, fairness, and alignment with human values.
- The effective contribution and involvement of Social Sciences and Humanities (SSH) disciplines and experts are required to produce meaningful societal effects.
Data Protection & Intellectual Property
- Open access to research outputs is a key requirement, following Horizon Europe's Open Science policy. If open access is not possible, a clear plan for how GPAI providers, policymakers, and other stakeholders will access the research outcomes must be provided.
- Intellectual property arising from the action and its results must remain within the recipient entity during and after completion of the action, without control or restrictions by non-eligible countries or entities.
Risk Management
- Projects must focus on proactively identifying and forecasting emergent capabilities in GPAI models, encompassing both beneficial and potentially detrimental uses.
- This includes assessing potential risks in critical areas such as chemical, biological, radiological, and nuclear (CBRN) hazards or cybersecurity threats.
Innovation & Quality
- Emphasis is placed on innovative methods for assessment and the generation of example benchmark tests.
- The developed frameworks, methodologies, and tools must be responsive to the concerns of all relevant stakeholders.
Partnerships & Ecosystem Development
- Projects are expected to implement the co-programmed European Partnership on AI, data and robotics (ADRA) and contribute to the GenAI4EU initiative, which is part of the broader Apply AI strategy.
- Collaboration and synergy development with existing projects and other relevant international, European, national, or regional initiatives (e.g., Digital Europe Programme) are encouraged.
Grants8