Artificial Intelligence is the New Astrology of Software Quality
This paper seeks to explore the limitations of Artificial Intelligence as a tool to automate decision making, including software testing.
Artificial Intelligence / Machine Learning are catchall terms, which often involve neural nets which are 'trained' using training data and other types of input in order to then respond to future input. A well trained AI is able to 'accurately' respond to stimuli in ways that meet business needs, cutting down on human intervention or complex rule based coding. However, just as modern science is based on fundamental first principles, software applications are supposed to be driven by clearly defined business rules and attendant business logic. Yet, AI frequently amounts to statistical correlation between 'black box' models and training data with unknown scientific, legal, moral or ethical assumptions. Therefore, concerns about biased and insufficient training data are only part of the problem.
When considering implementing AI or AI assisted technologies, management must understand the business risks of such a move, despite the seemingly unlimited enthusiasm and optimism of the technologists. The authors believe these considerations are especially relevant in the quality assurance of mission critical applications in large enterprises and government organizations." Management, Data Science, AI, Machine Learning, Business Rules / Business Logic
Jack McDowell, interim Statewide QA Program Manager, Enterprise Information Services - State of Oregon
Jack McDowell is an Operations & Policy Analyst for the State of Oregon's Statewide QA and E-Government Program. Before this, he was a web developer and the chief editor of a community
newspaper in Arlington, Virginia. He holds a Master's degree in political science from the University of Oregon and a certification in ITIL.
Ying Ki Kwong, E-Government Program Manager, Enterprise Information Services - State of Oregon
Ying Ki Kwong, PhD, PMP is the E-Government Program Manager in the Office of the State CIO in Oregon state government. Prior to this role, he was Statewide QA Program Manager, IT Investment Oversight Coordinator, and Project Office Manager of the Medicaid Management Information System Project.
In the private sector, Dr. Kwong was CEO of a Hong Kong-based internet B2B portal for trading commodities futures and metals. He was a program manager in the Video & Networking Division of Tektronix, responsible for worldwide applications & channels marketing for a line of video servers in broadcast television applications. In these roles, he has managed software-based systems / applications, products, and business process improvements.
He received the doctorate from the School of Applied & Engineering Physics at Cornell University and was adjunct faculty in the School of Business Administration at Portland State University. He holds certifications in PMP, ITIL, and IT Service Management. He is a board member of PNSQC.