Soon, all journalism will be carried out by robots — or at least, that was the fear among some journalists when major organizations like the Associated Press turned to automation technology to churn out some of their more rote news stories, like quarterly earnings reports and sports game recaps.

The AP is not the only news organization to lean on some sort of automation technology to power its journalism: newsrooms from The New York Times to ProPublica to Oregon Public Broadcasting have turned to algorithms as well.

Automated journalism is here to stay, according to a new report from Columbia’s Tow Center and research fellow Andreas Graefe. But what are the most efficient uses of this technology, and what are its limitations? What is it good for that news organizations might not have considered yet? How do readers respond to automated articles? How does the spread of automated journalism affect the way human journalists do their jobs? The report explores some of those issues in depth, presenting several case studies (and allaying some fears of a robot takeover).

Potentials and limitations

Automated journalism is still a fledgling field. It isn’t perfect, and it isn’t going to produce beautifully crafted sentences (any time soon). It won’t work in domains where no structured data is available or the data available is fuzzy, and it can’t provide the why of a story, only the what.

Algorithms for generating automated news follow a set of predefined rules and thus cannot innovate. Therefore, their application is limited to providing answers to clearly defined problems for which data are available. Furthermore, at least at the current stage, the quality of writing is limited.

But its potential in increasing speed, scale, and accuracy is great — algorithms “do not get tired or distracted.”

Automation allows for expanding the quantity of news by producing stories that were previously not covered due to limited resources. For example, both the Los Angeles Times and the Associated Press reported that automation increased the amount of published stories by more then ten times. Similarly, while human journalists have traditionally only covered earthquakes that exceeded a certain magnitude or left significant damage, [the Los Angeles Times’s] Quakebot provides comprehensive coverage of all earthquakes detected by seismographic sensors in Southern California. While any one of these articles may attract only a few hits in targeting a small audience, total traffic increases with positive effects on advertising revenues.

Key questions remain

The Tow Center report also outlines some implications of automated journalism as it applies to journalists, news organizations, readers, and even society at large. Human journalists are no match for the speed and scale allowed by algorithms, but journalists can turn their attention to building skills beyond writing speed.

New technologies need new masters, too: The Associated Press, for instance, has hired an automation editor to surface which areas of coverage can be automated.

Advocates of automated journalism argue that the technology benefits news consumers by providing new content that was previously unavailable and personalizes that content to meet the needs of the individual consumer. This raises two important questions. First, how do news consumers perceive the quality of automated news? Second, what are news consumers’ requirements regarding algorithmic transparency?

Like journalists, news consumers seem to find the quality of writing of automated content to be quite low. But we don’t know much about what else readers would like to know about such algorithmically generated content, or whether they want any more information at all.

The Tow report also brings up the ultimate question of how widespread reliance on algorithms might impact society:

If algorithms were employed for public interest journalism, questions will arise as to whether we can and should trust algorithms as a mechanism for providing checks and balances, identifying important issues, and establishing a common agenda for the democratic process of public opinion formation. Furthermore, future research will need to study the implications for democracy if algorithms are to take over journalism’s role as a watchdog for government.

The full report, which was partially funded by the Knight Foundation, is available here.