Artificial Intelligence is one of the fastest growing technologies of the21st century and accompanies us in our daily lives when interacting with applications. However, reliance on such technical systems is crucial their widespread applicability and acceptance. The societal tools to reliance are usually formalized by lawful regulations, i.e., standards,, accreditations, and certificates. Therefore, the T\\"UV AUSTRIA Group in with the Institute for Machine Learning at the Johannes Kepler Linz, proposes a certification process and an audit catalog for Learning applications. We are convinced that our approach can serve as foundation for the certification of applications that use Machine Learning Deep Learning, the techniques that drive the current revolution in Intelligence. While certain high-risk areas, such as fully robots in workspaces shared with humans, are still some time away certification, we aim to cover low-risk applications with our procedure. Our holistic approach attempts to analyze Machine applications from multiple perspectives to evaluate and verify the of secure software development, functional requirements, data quality, protection, and ethics. Inspired by existing work, we introduce four levels to map the criticality of a Machine Learning application the impact of its decisions on people, environment, and. Currently, the audit catalog can be applied to low-risk within the scope of supervised learning as commonly encountered in. Guided by field experience, scientific developments, and market, the audit catalog will be extended and modified accordingly.
No takes yet. Share an insight, caveat, or question.
Winter et al. (2021) studied this question.