Enhancing the Security, Privacy and Robustness of AI Models and Systems (SecureAI)

The EU funding call “Enhancing the Security, Privacy and Robustness of AI Models and Systems (SecureAI)” targets research institutions, universities, companies, and start-ups in Europe developing innovative AI solutions in safety-critical areas. With a total budget of approximately EUR 21 million, the Horizon Europe programme will fund research and innovation projects (Innovation Actions) from 3 March 2026 until 15 September 2026 aimed at increasing resilience against adversarial attacks, data poisoning, and backdoors. The funding rate is up to 70% of eligible costs (personnel, materials, travel, and subcontractor expenses). Eligible participants are legal entities from EU Member States and associated countries forming consortia with at least three independent partners. A prerequisite is exemption from high-risk suppliers according to the EU 5G Toolbox as well as a convincing concept for data and system security. Evaluation criteria include the degree of innovation, expected societal impact, and consortium quality.

The focus is on methods for developing AI models that employ privacy-enhancing technologies throughout the entire AI lifecycle, such as federated learning, homomorphic encryption, and secure inference. Furthermore, automated procedures for detecting manipulated training data and robust optimization approaches are to be developed. The projects should have an impact on critical infrastructures, public administrations, and industrial production processes and actively contribute to the implementation of the EU AI Regulation. For the application, the standard application form, a detailed project description, the financing plan, and a declaration on security issues must be submitted. The results of the call are expected to be announced within five months after submission, and the agreement can then be concluded promptly.