Chapter 79. Make Accountability a Priority

Yiannis Kanellopoulos

There is little doubt that algorithmic systems are making decisions that have a great impact on our daily lives. As Yuval Noah Harari notes in his book 21 Lessons for the 21st Century (Random House), “Already today, ‘truth’ is defined by the top results of the Google search.” So transparency about the function of these systems matters not as an end in itself but merely as a means toward accountability.

According to assistant professor Nicholas Diakopoulos, director of the Computational Journalism Lab (CJL) at Northwestern University, accountability in this context means the degree to which one decides when and how an algorithmic system should be guided (or restrained) in the risk of crucial or expensive errors, discrimination, unfair denials, or censorship.

Simply put, holding a system accountable means we should control it at a technical as well as an organizational level. This is important, especially if we consider (a bit simplistically) that an algorithmic system is nothing more than a piece of software that:

  • Solves a business problem set by the organization that procures it (the system)

  • Receives data as input that has been selected and most likely preprocessed either by a human or an automated process

  • Utilizes a model (e.g., support vector machine, deep learning, random forest, and others) ...

Get 97 Things About Ethics Everyone in Data Science Should Know now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.