TÜV AUSTRIA presents the Trusted AI Framework, a pioneering approach to the assessment and certification of artificial intelligence systems used in safety-critical applications. In cooperation with the Johannes Kepler University Linz (JKU), the Software Competence Center Hagenberg (SCCH) and the TRUSTIFAI joint venture, an end-to-end audit catalog has been developed to translate technical, legal and ethical requirements into verifiable criteria.
TÜV AUSTRIA’s TRUSTED AI framework is based on three central pillars: secure software development, functional reliability requirements, and ethical and data protection guidelines. It offers a structured methodology for assessing the functional reliability of machine learning systems – including risk-based minimum requirements and statistical tests with independent data samples.
“With the TRUSTED AI framework, we are for the first time creating the basis for transparent and reproducible certification of AI systems – a decisive step towards their safe use in critical areas such as medicine, industry or mobility,” explains Andreas Gruber, Managing Director of TÜV AUSTRIA’s TRUSTIFAI joint venture.
Target group: those responsible for implementing the requirements of the EU AI Regulation
TÜV AUSTRIA’s white paper is aimed specifically at those responsible for system development, corporate governance, quality and risk management. “These people are responsible for implementing regulatory requirements and therefore play a key role in the successful implementation of AI,” Gruber continues.
The white paper presents not only the methodology, but also the practical experience gained from applying the test catalog. It highlights typical sources of error such as data leakage, inadequate domain definitions or insufficient testing of AI applications, and offers specific recommendations for action.
“Our experience from numerous customer projects shows that companies need concrete requirements and implementation aids for the development and operation of AI systems. As part of our research cooperation with JKU and SCCH, our white paper offers a comprehensive overview of the technical and organizational processes, procedures and measures that should be implemented as the state of the art,” says Stefan Haas, CEO of TÜV AUSTRIA.
With the TRUSTED AI framework, TÜV AUSTRIA provides a practical roadmap for suppliers, users and regulatory authorities – towards legally compliant, functionally reliable and certifiable AI systems in line with European standards.
“AI regulation must go beyond mere principles – it must be verifiable in practice. Our framework combines technical best practice with the requirements of the European AI Regulation and makes trust measurable,” Andreas Gruber emphasizes.
Check certificate
I am looking for a solution but do not know its title.
