Dependable Systems and Software
Dependable Systems and Software
The CARAT project investigates how certification can serve as a pivotal regulatory tool to manage risks and ensure trustworthiness in Artificial Intelligence (AI) systems. Given the rapid expansion and integration of AI across various industries, the initiative focuses primarily on Germany, Europe, and Brazil, evaluating how effective certification processes can mitigate critical AI-related risks. These risks include discrimination, transparency deficits, privacy violations, safety concerns, and ethical considerations. The overarching aim is to foster the development of a comprehensive legal and technical framework that promotes compliance, innovation adaptability, and international harmonization of AI governance standards.
The expected outcomes of CARAT include comprehensive legal assessments and practical proposals for voluntary and mandatory certification frameworks, along with clear technical guidelines tailored to meet legal compliance requirements. Findings from the project will be shared broadly through international symposia and consolidated into a final, comprehensive publication aimed at policymakers, industry stakeholders, academic researchers, and the wider public.
More info (in German)