Three new projects, ROBUSTIFAI, AIXPERT, and TRUMAN, aim to foster better trust, safety, and transparency in artificial intelligence across Europe
To make responsible and human-centric artificial intelligence (AI), the European Union has launched three new research projects, ROBUSTIFAI, AIXPERT, and TRUMAN, under the Horizon Europe programme.
Coordinated by the European Health and Digital Executive Agency (HaDEA), these projects are supported by a total EU contribution of €21.4 million and fall under Horizon Europe Cluster 4: Digital, Industry and Space.
These projects aim to enhance the robustness, transparency, and explainability of AI systems, aligning with key European values such as fairness, accountability, and human oversight.
ROBUSTIFAI: Human-centric generative AI
Generative AI (GenAI) has changed how machines understand and create content, but it still faces challenges related to robustness and reliability. The ROBUSTIFAI project hopes to address these limitations by integrating human needs and symbolic reasoning into the design of GenAI systems.
By combining neural network models with symbolic methods and focusing on adaptability to various users and environments, ROBUSTIFAI aims to make GenAI more robust from a technical, operational, and user perspective. The project’s approach will be tested in real-world use cases across three distinct sectors: automotive, service robotics, and cybersecurity.
AIXPERT
AIXPERT introduces a multi-agent, GenAI-powered platform designed to make AI systems more transparent, explainable, and accountable. This flexible platform accommodates a variety of AI models, allowing for consistent oversight and user feedback regardless of the underlying architecture.
The platform’s adaptability makes it suitable for a wide range of industries. The project will explore practical applications in healthcare, recruitment, manufacturing, education, and the creative sector. By embedding real-time feedback and multi-modal data analysis, AIXPERT supports the creation of user-friendly, transparent AI systems that meet high ethical standards.
TRUMAN: Building rust in AI systems
The TRUMAN project focuses on building the resilience of AI systems against risks related to security, privacy, and fairness. Taking a life-cycle approach, TRUMAN addresses vulnerabilities from data collection through to system deployment.
The project aims to develop tools that can detect and defend against both existing and emerging threats, while also emphasising the need for transparent communication about AI technologies and their associated risks. TRUMAN will involve users in evaluating and improving these systems, ensuring that human trust remains at the core of AI development.
To validate its solutions, TRUMAN will conduct tests in marketing, IT security, finance, and healthcare sectors where trust in AI is particularly critical.
Supporting EU AI plans and community collaboration.
All three projects will actively contribute to EU AI initiatives, including the AI-BOOST project and the AI-on-Demand platform. By doing this, they strengthen the European Commission’s commitment to building a robust, collaborative AI ecosystem that strikes a balance between innovation and ethical standards.
With a strong focus on real-world applications, these projects are set to play a crucial role in shaping the future of trustworthy AI in Europe. They show a strategic investment in developing technologies that not only advance the state of the art but also ensure AI systems are safe, explainable, and aligned with societal values.
For more on research priorities and upcoming funding calls, visit the HaDEA Call for Proposals and the Horizon Europe Work Programme 2025.
