Fair, Responsible, Secure and Private – The Changing Dimensions of AI
Year: 2020
Type: preprint
Abstract: There has been an ever-increasing footprint of AI and machine learning in our business value chain and processes. Many process automations that consume machine leaning algorithms are being rolled with the aim of reaching the last mile of automation. Automations are getting smarter, learning from user actions, predicting business decisions and recommending choices. With all the evident positives, there is a question being asked as to how can we be sure that these algorithms are taking or recommending the right decision?
Leading banking, life science and healthcare corporations are already invested into policy formulation towards controlling AI and machine learning algorithms-based programs with expectations of conformance to these policies.
Accuracy cannot be the only parameter to evaluate effectiveness of machine learning algorithms simply because algorithms learn based on training but cannot identify if the training data itself is faulty.
This paper evaluates methods that can be used to make machine learning based systems fair, responsible and ethical.
Source: SSRN Electronic Journal
Author: Nirmalya Shome
Institution: Tata Consultancy Services (India)
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3156273456