Imagine buying a car without knowing if the brakes actually work or if the airbags will deploy. Many companies today use AI like that—they know it's fast, but they aren't sure if it's safe or fair. This work creates a digital 'safety inspection' checklist to prove an AI system is reliable before it's used in the real world. It's like a quality seal for software that handles critical tasks.
AI developers often prioritize speed and performance over safety and reliability. This leads to a lack of trust from users and potential legal risks when deploying AI in critical sectors like health or transport.
A digital assessment tool containing over 50 metrics to measure AI security, transparency, and robustness, validated through seven real-world pilots.
If you are a medical AI software developer dealing with strict patient safety rules — this project developed a set of over 50 metrics that prove your system is secure and usable. This helps you get your product approved for clinical use faster.
If you are an autonomous vehicle manufacturer dealing with unpredictable road risks — this project developed a risk management system that tests for robustness and traceability. This reduces the chance of costly system failures in the field.
If you are a robotics integrator dealing with human-robot collaboration safety — this project developed a way to measure and optimize trust in critical domains. This ensures your robots are accepted by workers and meet safety standards.
Based on available project data, no pricing or licensing costs are mentioned as the project is EU-funded.
Yes, the project is specifically designed around seven international Large Scale Pilots in critical domains to ensure the tools work in real-world settings.
Based on available project data, specific IP and licensing terms are not provided, though it aims to support compliance with the EU AI Act.
The tools are designed to help companies comply with emerging laws like the EU AI Act and follow ENISA guidelines.
The project runs from 2024-01-01 to 2028-03-31, with initial versions of the assessment tools already released during the first reporting period.
The project is backed by a diverse group of 18 partners across 9 countries. With a 22% industry ratio (4 companies, including 4 SMEs), the consortium balances academic research from 5 universities and 5 research centers with practical application. This mix ensures that the 50+ metrics developed are not just theoretical but are tested by actual industry players in critical sectors.
Contact IDRYMA TECHNOLOGIAS KAI EREVNAS in Greece
Contact us to find out how to implement the FAITH metrics in your AI pipeline.
