Intellimetry as a way to ensure AI trustworthiness

Garbuk S.

The article is about the issues of AI systems compliance confirmation with the requirements. It is shown that in the conditions of using the artificial intelligence technologies for solving various application tasks, the trustworthiness of AI systems on demand is the particular importance. Features of artificial intelligence systems trustworthiness confirmation are considered. A set of measuring methods for the functionality and quality assessment of intelligent systems is proposed to allocate in an independent direction, called "Applied Intellimetry". The formalization of basic concepts and problems of Applied Intellimetry are proposed.

Keywords: AI systems quality, AI systems trustworthiness, applied AI tasks, Applied Intellimetry

In book

The Proceedings of the 2018 International Conference on Artificial Intelligence: Applications and Innovations. Limassol, October 8-12.

[б.и.], 2018.