loading page

Standardization on Bias in Artificial Intelligence as Industry Support
  • Ewelina Szczekocka ,
  • Christele Tarnec ,
  • Janusz Pieczerak
Ewelina Szczekocka
Author Profile
Christele Tarnec
Author Profile
Janusz Pieczerak
Author Profile

Abstract

Industry strives to have trustworthy AI systems, then to recognize and implement Responsible AI principles to achieve that goal. It is important for us to find solutions that will help implement trustworthy AI systems. Standardization activities are an important element for us in this regard, as they provide a platform for discussion for industry towards use and help work out practical rules and requirements. This article first provides an overview of the bias issue in Artificial Intelligence, and presents the concept of Responsible AI. Second, we present examples of interesting ongoing works by international  standardization organizations on bias in AI systems. Then we identify a gap between the principles defined by high-level legal studies on responsible AI, including the European AI Act proposal, and  practical implementations of these principles. Finally we state how international standardization may fill this gap. And perhaps the travel still to conduct to fill it?