Send message to Author

Short abstract:

This paper explores how fact checkers enthusiastically take part in automating specific tasks in their job, ones they deem as repetitive and boring. Based on extensive field work, it shows how these automation attempts reshape the boundaries between technology and humans in journalistic work.

Long abstract:

Fact checkers have always been a relatively technologically savvy professional group within journalism. Now, they are standing at the forefront of the efforts to infuse AI into their work.

The talk is based on 3 years of field work and interviews with fact checkers around the world who take part in building tools to automate their work. Differently from doomsday scenarios of “robots stealing our jobs”, fact checkers mostly accept and welcome automation of what they see as boring and repetitive tasks, in favor of providing them with more time to engage in difficult tasks that they do not want machines to conduct.

These encounters draw the lines of what they see as inherently human and intellectually stimulating tasks, that must remain explainable and transparent, and what they see as a waste of human capacities, thus reshaping the boundaries between human and machine in journalistic work.

In addition, if so far in journalism mostly audience behaviour was datafied, automation of fact checking takes the datafication to a new realm - the journalistic work routines. The automation is based on training data that is the result of fact checkers past work, and at the same time building an automated tool provides a framework for capturing future actions - their outputs, what they write, link, search and decide - and turning them into training data. The machine learning tools built by them help exacerbate the datafication and deepen it.