RobustifAI (Horizon Europe project)

active

A three-year Horizon Europe consortium (budget €9.3M) launched in June 2025 to develop rigorous, human-centric methods for reliable and trustworthy generative AI in human cyber-physical systems, with Hebrew University’s contribution led by Prof. Guy Katz.

Endorsements support Hebrew University of Jerusalem.

People

Guy Katz

Principal Investigator for Hebrew University’s RobustifAI team

Funding Details

Start Date: Jun 1, 2025
End Date: -
Expected Duration: -
Funding Raised to Date: -
Annual Budget: -
Monthly Burn Rate: -
Current Runway: -
Funding Goal: -
Funding Stage: -
Fiscal Sponsor: -

Project Details

RobustifAI is a Horizon Europe project aimed at creating a comprehensive design and deployment methodology for reliable, robust, and trustworthy generative AI. Officially launched on 1 June 2025 with a total budget of €9.3 million and a planned duration of 36 months, the consortium includes 18 partners from 12 countries: seven universities, two research and technology organisations, five SMEs, and four major industrial companies. The project focuses on human cyber‑physical systems such as autonomous vehicles, healthcare, energy, and smart manufacturing, where safety and robustness are mission‑critical. Hebrew University is one of the university partners, and Professor Guy Katz of its School of Computer Science and Engineering leads the university’s contribution. Drawing on his work in formal verification of systems with machine‑learned components, the Hebrew University team works on methods to analyze and strengthen the robustness and safety of foundation models and other GenAI components deployed in these environments.

Theory of Change

RobustifAI assumes that deploying generative AI in safety‑critical human cyber‑physical systems requires formally grounded methods that can guarantee robustness and trustworthiness. By bringing together universities, research institutes, SMEs, and major industrial partners to integrate neural and symbolic techniques with formal verification and human‑centric design, the project aims to produce tools, methodologies, and case studies that industry and regulators can adopt. These outputs are intended to reduce failure modes and vulnerabilities in foundation models used in domains such as transportation, healthcare, and manufacturing, thereby enabling safer large‑scale deployment of generative AI.

Grants Received– no grants recorded

Discussion

No comments yet. Be the first to share your thoughts.