Socio-technical research

Many engineering systems have traditionally been dominated by mechanical and electrical engineering however current trends have seen the influx of advanced computing and AI systems. AI is associated with its ethical issues. From autonomous vehicles to health diagnostic systems, the infiltration of machine learning AI demands a new approach to product development. Hence traditional quality management approach needs to be improvised. My research here focuses on the socio-technological approach to improving designs. One research question is how to incorporate fairness early in the design process to avoid the risk of bias introduced by AI. 

  I am currently finalizing a study on how machine learning models discriminate. Here, I explore the limits of representative data and other technical aspects in eliminating discrimination in algorithms. I explain how models are formulated and how they can consciously or unconsciously incorporate bias. I present a procedural audit perspective that considers socio-technical factors and a proper problem definition. The study is tentatively planned for submission to a reputable technology ethics journal.

Another current research I am undertaking is 'algorithm that audits algorithms'. I am proposing a regression model approach to auditing machine learning predictions and classifications.