A study led by Neil Thurman explores journalists’ perceptions of automated journalism in terms of its quality and ethical implications. Automated journalism refers to “algorithmic processes that convert data into narrative news texts with limited to no human intervention beyond the initial programming” (Carlson 2014: 416).
Automated journalism has clear potential for speeding up journalistic routines and reducing the costs of multi-platform publishing. In particular, it seems valuable for generated simple news reports in areas such as sports and business and for adapting articles for different devices and platforms. However, journalists are understandably concerned about the possibility of being replaced by algorithms. In addition, Thurman’s study points to concerns about the quality of automated journalism.
To gauge how journalists perceive the implications of automated journalism, the researchers interviewed ten journalists and editors from CNN, BBC, Thomson Reuters, Trinity Mirror, and The Sun. The journalists’ opinions were gathered following a period of training and hands on experience with automated software. Journalists concluded that automated reporting “lacks the context, complexity, and creativity of traditional reporting”.Six main limitations of were identified:
1.Its reliance on single, isolated data streams.
2.The one-dimensional nature of the quantitative data feeds it relies on.
3.The difficulties of interrogating that data.
4.The lack of human angles in the texts generated.
5.The requirement to template stories—predicting “top lines”—in advance.
6.The difficulty of working creatively with the data in the templating process.
A journalist at Thomson Reuters pointed to an automated story about European Central Bank (ECB) interest rates and noted the absence of any “semantic” understanding of how the ECB president discussed the announcement. The data journalist at Trinity Mirror agreed that “there’s always going to be things within data feeds that will stand out more to a human than it will to a computer”. Similarly, respondents highlighted a story about the collapse of BHS and noted that data alone gives no context to the “flamboyant lifestyle and arrogance” of the former owner, Philip Green.
Some journalists found the software’s functionality frustrating, particularly regarding the flexibility required for investigative styles of data journalism. Although respondents did believe the technology could usefully augment their work, they found automated output to be “throw away, repetitive, not particularly interesting”. The requirement to template automated stories in advance was perceived as a significant drawback. A BBC journalist noted “how limiting writing everything in advance is. You can’t get a reaction to those numbers, you can’t explain or interrogate them, because you wrote it all before the numbers came out.”
Participants discussed a range of ethical issues including transparency and bias as well as wider societal concerns. A BBC news reporter highlighted the complexity of determining authorship when “bots are part human and part bot”. If automated tools are “written by a human being in the first place”, how autonomous is automated news? Although an algorithm is less likely to make simple mechanical errors, there was concern that assumptions about the reliability of automation would make verification less likely and oversight harder.
Although the journalists’ responses were largely negative, many did appreciate the potential of automation as a supplement to traditional journalism and many recognised the capacity of the technology to improve upon its current limitations. Within the technology world, the roll out of chat bots and virtual assistants is part of the general move away from keyboards, screens and even phones, which could bring further disruption to the routines and practices of journalistic work.
Read the full article (pdf).
Subscribe to FuJo’s newsletter.
