AI quality becomes verifiable: MISSION AI presents quality standard and digital testing portal for low-risk AI
MISSION AI is presenting a voluntary quality standard for low-risk AI systems today at AI Quality & Governance Day in Berlin. In addition, a digital testing portal is being launched to facilitate practical application for companies. Together, the standard and portal enable a structured assessment of the quality of AI systems that fall below the high-risk threshold. For this area, the EU AI Act currently only stipulates transparency requirements.
The MISSION AI quality standard summarizes key principles of trustworthy AI – including transparency, non-discrimination, and reliability – in six quality dimensions. These have been translated into clear criteria and verifiable measures and form the basis of a self-administered testing procedure. This procedure guides companies step by step through the evaluation of their AI system. A protection needs analysis determines which requirements are the focus in the respective application context. The final test report shows the degree of compliance with the quality requirements and highlights areas where AI systems can be specifically improved.
The standard offers several advantages for companies: AI providers receive structured proof of quality that can serve as a basis for dialogue with investors, customers, and regulatory stakeholders. Start-ups and SMEs in particular can benefit from the opportunity to systematically document quality measures and demonstrate them in procurement and tendering procedures. For public clients and procurement agencies, uniform criteria are created on the basis of which AI solutions can be compared in a structured manner. At the same time, the standard supports companies in addressing the requirements of the EU AI Act early and systematically.
Manfred Rauhmeier, Executive Board Member of the acatech Foundation:
"Standards are crucial for building trust in the application of new technologies. With its quality standard, MISSION KI offers companies valuable support so that they can make their AI applications trustworthy from the outset. The new standard is widely applicable. It creates transparency and provides reliable proof of quality to customers and partners. In this way, we are strengthening the competitiveness of all players in our economy."
MISSION KI has developed the standard over the past two years in collaboration with PwC Germany, TÜV AI.Lab, the VDE Association for Electrical, Electronic & Information Technologies, the AI Quality & Testing Hub, and the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS. The standard thus combines expertise from standardization, testing, consulting, and research.
The standard was tested with partners such as KfW, CLAAS, ZEISS, enamentis, and BeIntelli using real AI systems. The findings from these tests have been incorporated into further development. An independent group of experts from business, standardization, and research has continuously accompanied the development process.
Hendrik Reese, PwC EMEA Responsible AI Leader:
"Voluntary testing according to the MISSION AI quality standard is an important step toward strengthening trust in AI beyond high-risk applications. For the first time, companies now have a tool that enables them to evaluate their AI systems according to specific criteria – in a practical, comprehensible manner that is compatible with regulatory developments such as the EU AI Act. This makes responsible AI a competitive factor."
The audit portal is now available in a beta version and digitally maps the audit process. Guided processes and integrated assistance make self-auditing easier and more efficient for companies. In the future, the portal will be expanded to include automated audit steps and interfaces to accredited testing laboratories.
More information about the quality standard and testing portal here
