AI Assurance Processes

9 Pages Posted: 20 Oct 2020

See all articles by Emre Kazim

Emre Kazim

Holistic AI; University College London

Adriano Koshiyama

Department of Computer Science, University College London

Date Written: September 2, 2020

Abstract

The ethics of Artificial Intelligence (AI Ethics) can be thought of as undergoing three broad phases, with the first two being principles and processes, to the current phase which we read in terms of assurance. In this article we anticipate and expand upon themes in the AI assurance space: namely; i. general and sector specific assurance, ii. governance via technical and impact assessments, iii. the use of ‘traffic-light’ user friendly monitoring interfaces, iv. risk assessments and ‘red teaming’ in order to mitigate unknown risks, v. certification, and vi. insurance of AI systems. The article concludes by noting that the manner in which AI assurance matures is likely to be as a result of the extent and reach of any regulatory intervention.

Keywords: Artificial Intelligence, AI Ethics, Assurance, Audit, Impact Assessment, Certification, Red Teaming, Insurance, Regulation, Standards

Suggested Citation

Kazim, Emre and Koshiyama, Adriano, AI Assurance Processes (September 2, 2020). Available at SSRN: https://ssrn.com/abstract=3685087 or http://dx.doi.org/10.2139/ssrn.3685087

Emre Kazim (Contact Author)

Holistic AI ( email )

18 Soho Square
London, W1D 3QH

University College London ( email )

United Kingdom

Adriano Koshiyama

Department of Computer Science, University College London ( email )

Gower Street
London, London WC1E 6BT
United Kingdom

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
485
Abstract Views
1,790
Rank
107,917
PlumX Metrics