The Conundrum of Machine Learning and Cognitive Biases

Posted: 19 December 2016 | By Darcie Thompson-Fields

Machine learning is on the rise due to the technological convergence of the growth of big data, decreasing data storage costs, increasing computing power, improved artificial intelligence algorithms and acceleration of cloud computing. Machine learning is the ability for computers to learn without explicit programming. It’s analogous to the human ability to identify an octopus based on the set of data input that goes to the brain, such as eight arms, tentacles, lack of skeleton and other characteristics, without having prior knowledge of every type of cephalopod mollusk in existence.

However, human decision-making is subject to numerous cognitive biases that can easily distort judgement. For example, iconoclastic author Tom Peters highlights 159 cognitive biases that impact management decision-making (Peters, Tom. “159 Cognitive Biases Between You and Good Judgment (Good Luck!).” 29 May 2014). Emotions such as anxiety, fear or anger could easily cloud a person’s judgement. Human thinking is prone to the use of cognitive heuristics, a shortcut that may lead to biases and faulty decisions.

Given a computer is devoid of emotion and the hubris of human ego, it would seem logical that machine learning is not impacted by cognitive bias. The caveat is that humans are creating and monitoring the programming for machine learning. The areas of cognitive bias vulnerability for machine learning include:

  • Data structure, collection and sources
  • Data set size
  • Level of objectivity in the data
  • Weight assignments to data points
  • The absence or inclusion of indicators
  • The inherent cognitive biases of the human programmer

Machine learning technology is deployed today for many business uses, including self-driving cars, online recommendation, search engines, handwriting recognition, computer vision, online ad serving, pricing, prediction of equipment failure, credit scoring, fraud detection, OCR (optical character recognition), spam filtering and many other uses. The growing ubiquity of machine learning in business makes it critical to mitigate the introduction of human cognitive biases into the machine.

More articles written by Cami Rosso: 

The Importance of Philosophy in the Future of AI and Humanity

Why Artificial Intelligence is the Next Revolution 

If you would like to connect with Cami, you can do so on LinkedIn and following her on Twitter @CamiHRosso

Cami has worked in Silicon Valley, Wall Street and globally for leading companies; Apple, PepsiCo, Macromedia, Morgan Stanley, Oracle and Adobe, to name a few.