Fill out the form
Horizon Europe programme
Official call identifier: HORIZON-CL3-2026-02-CS-ECCC-02
Objective of the call
The call aims to strengthen the resilience, security, and privacy of AI models and systems against adversarial threats. It focuses on developing robust AI solutions capable of resisting manipulation, ensuring data confidentiality, and enabling secure and trustworthy AI deployment in sensitive environments.
Scope of activities
Proposals should address the security and robustness of AI systems against adversarial attacks and vulnerabilities, including:
- Development of robust AI models resistant to adversarial manipulation through techniques such as adversarial training and robust optimisation
- Detection and mitigation of attacks including data poisoning, backdoor exploitation, and misclassification
- Development of real-time anomaly detection and mitigation mechanisms
- Improvement of detection and validation of manipulated or compromised training data
- Development of privacy-preserving AI systems, including federated learning, secure aggregation, and encrypted computation
- Implementation of trusted AI deployment environments (e.g. on-premise or regulated cloud) ensuring compliance with regulatory frameworks such as the AI Act
Some activities may involve classified background or produce security-sensitive results. Participation is limited to legal entities established in EU Member States and Associated Countries. Entities controlled by non-eligible countries are not allowed to participate.
Eligible applicants
- Innovative SMEs, startups and enterprises aiming to scale and strengthen competitiveness
- Public-sector organisations driving transformation and societal impact
- Research and academic institutions commercialising knowledge
- Non-profit organisations delivering purpose-driven innovation
- Organisations based in EU Member States or Associated Countries (described in General Annex A, General Annex B, General Annex C)
- Minimum 3 independent partners from 3 eligible countries
Specific consortium we are looking for:
- AI researchers, cybersecurity experts, data scientists, and organisations specialised in privacy-enhancing technologies, federated learning, and secure AI deployment.
- Participation of entities with expertise in regulatory compliance, AI validation, and secure infrastructure environments is expected.
Eligible costs
Eligible costs will take the form of a lump sum contribution under the Horizon Europe Lump Sum Model Grant Agreement. Activities may include:
- Research and development of cryptographic primitives and protocols
- Software development and implementation
- Formal verification and tool development
- Integration into software workflows
- Open-source library development
- Testing, validation and performance evaluation
- Security and compliance activities
Funding conditions
Type of action:
Innovation Actions (IA)
Funding rate:
Up to 70% of total eligible costs (100% for non-profit legal entities)
EU contribution per project:
Around EUR 3,000,000 - 4,000,000
Number of funded projects:
Up to 5 projects
Project duration:
Typically 36 months
Deadline for submission
15 September 2026