Health information technology innovations drive us to clarify what machines and nurses each do best and robotic journalism is no exception. Robotic journalism (sometimes referred to as robo journalism or bot journalism) is the use of software algorithms designed to generate written articles. This automated journalism software goes beyond the tasks of computer searching, data mining, fact-checking, and data visualization to actually create content in a narrative text format. Researchers refer to this process of data-to-text as "natural language generation." A number of narrative content generation tools are now commercially available. Software that can write meaningful sentences is used for routine stories in fields where summarizing structured quantifiable data that are typically found in spreadsheets and databases are key content such as reports on weather, financial earnings, and sports.
Robotic journalism gained public attention when the Los Angeles Times published an Associated Press earthquake story in less than 3 minutes after the earthquake occurred (Schwencke, 2014). The journalism algorithm is programmed to code data into a template (style guide) using the same stock phrases that a reporter would use in creating an article. Automated algorithms perform simple, formulaic, repetitive writing tasks, which humans find undesirable, with greater speed and accuracy. News centers can distribute these articles in real-time to electronic news feeds. When a software program is first implemented, all the articles are reviewed by humans. When the software demonstrates success, articles are published automatically and only spot-checked.
Little is known about how readers recognize or respond to computer-generated articles. One study found readers perceived software articles to be boring but nonetheless objective and similar to journalist articles (Clerwall, 2014). The software-generated article may identify the author as a software algorithm and identify the data source in the byline. Some articles about robot journalism have added a byline stating that "no robots were involved in creating this content." Clearly, we need to begin reading bylines carefully.
Journalism software is designed to assist, not replace, human journalists. Dual-degree programs are now preparing programmer-journalists and many journalists see automation as an opportunity to be more creative. Routine and tedious tasks can be delegated to bots, leaving journalists to focus on the creative human aspects. Humans are still needed to review the software article and provide a "write-through" or the editing and refining of the automated story. Humans are still needed to identify important issues, make editorial decisions, and uphold journalism quality. Robotic journalism raises some ethical issues that humans will need to negotiate, such as data source authority, source disclosure, and authorship.
Although healthcare is a field where summarized structured quantifiable data are key content, use of robotic journalism is not reported. One blog post ventured to anticipate future use with a fictitious scenario, where an algorithm called an "RTC bot" wrote a summary of a clinical trial using the trial database as the data source (Smith, 2014).
New journalism technology may soon challenge nurse authors and editors to incorporate robot-generated content. Success will require leveraging the nursing journalistic expertise that computers do not have, such as professional wisdom, linguistically skilled writing, analytic skills, creativity, curiosity, passion, and genuine personal relationships with a journalism team. We here at MCN have known this human nursing journalistic expertise in our colleague, Dr. Margaret Comerford Freda, Editor Emeritus, and we are clear on how she did it best. She will be sorely missed.
References