The process of gaining an AI Legibility Mark

Pillling, Franziska, Akmal, Haider, Coulton, Paul ORCID: https://orcid.org/0000-0001-5938-4393 and Lindley, Joseph ORCID: https://orcid.org/0000-0002-5527-3028 (2020) The process of gaining an AI Legibility Mark. In: CHI '20: CHI Conference on Human Factors in Computing Systems, 25 April 2020 - 30 April 2020, Honolulu, Hawaii, USA.

Preview

Accepted Version Available under License In Copyright. Download (422kB) | Preview

Abstract

Researchers and designers working in industrial sectors seeking to incorporate Artificial Intelligence (AI) technology, will be aware of the emerging International Organisation for AI Legibility (IOAIL). IOAIL was established to overcome the eruption of obscure AI technology. One of the primary goals of IOAIL is the development of a proficient certification body providing knowledge to users regarding the AI technology they are being exposed to. To this end IOAIL produced a system of standardised icons for attaching to products and systems to indicate both the presence of AI and to increase the legibility of that AI's attributes. Whilst the process of certification is voluntary it is becoming a mark of trust, enhancing the usability and acceptability of AI-infused products through improved legibility. In this paper we present our experience of seeking certification for a locally implemented AI security system, highlighting the issues generated for those seeking to adopt such certification.