AI Case Study

Artificial Intelligence and News Bots: Journalists’ Responsibility to Harness Technology

Journalists can utilize the power of AI and news bots to communicate truthful information online and connect with their readers.

During the COVID-19 pandemic, misinformation has frequently been spread through word of mouth, incorrect news reports and false facts in social media banter. There is another source responsible for pandemic misinformation, however, that is often ignored: Twitter bots. In May of 2020, NPR reported that “nearly half of the Twitter accounts spreading messages on the social media platform about the coronavirus pandemic are likely bots,” according to researchers at Carnegie Mellon University (Allyn). These tweets included false information such as mannequins filling up hospital beds for COVID-19 patients and COVID-19 being spread through 5G wireless towers (Allyn). A bot can be defined as “a computer that attempts to talk to humans through technology that was designed for humans to talk to humans” (Roberts). Technology has leveled the playing field for journalists, citizens and robots to share information the same way. Social media has created an environment where the most prominent information isn’t necessarily a voice of authority, rather, users see information that is trending. Tweets and Facebook posts that are trending can sometimes be false information, which is one of the ways misinformation is spread through bots. This has left some journalists worried that the bot’s voices of misinformation could begin to speak louder than credible news sources (Zelizer 7). Journalists have the responsibility to harness AI and bots to communicate accurate and truthful information on social media and in their writing. 

Harnessing AI for journalistic writing has been a practice for the past ten years and has become an integral part in producing news. Automated journalism utilizes technology to produce news articles in seconds, expediting the journalistic process for news organizations. Journalists use artificial intelligence (AI) to create articles that are written by machines, rather than human writers. These machines are able to deliver articles at a rate of up to 2,000 articles per second and rely on natural language generation (NLG) to create a human-sounding narrative (Pressman). News organizations use software tools like Automated Insights, Narrative Science and United Robots to write articles ranging from sports to quarterly earnings to real estate market reports (Pressman). 

One of the first significant uses of AI in journalism was for writing articles about earthquakes. In March of 2014, the Los Angeles Times published an article about a 4.4 magnitude earthquake in the Los Angeles area eight minutes after it occurred. The LA Times was one of the first outlets to report the earthquake online because it was written by a piece of artificial intelligence called Quakebot (Levenson). Ken Schwencke, a database producer at the LA Times created a Quakebot to automatically report on earthquakes. The Quakebot takes an earthquake over 3.0 magnitude reported from the US Geological Survey, produces text, adds a map and writes a headline for the story (Levenson). Schwencke was positive about the new technological integration into reporting, saying, “it’s been a very helpful tool to get up something as soon as possible and let people focus on what’s actually happening outside of the basic information” (Levenson).

Ken Schwencke, a database producer at the LA Times created a Quakebot to automatically report on earthquakes. Courtesy The Atlantic. 

            Journalists have continued to use AI since then to report on quantitative information that is easier to acquire through the assistance of technology. Sometimes, journalists are even able to cater the quantitative information in a news article to the reader’s interest. In May of 2015, The New York Times published an article about the upward income mobility of children who grow up in places of poverty. Accompanying the article is an interactive map that uses the reader’s current location to show information about income mobility about their area. The AI then adapts to the reader using pre-assigned blocks of text, editing information based on the available data from the reader’s their current location (Lecompte). When the article was released, “many users didn’t notice that the text was assembled algorithmically. They just arrived at the page and thought their version of the story was the only version of the story” (Lecompte). This algorithmic method of catering articles to the reader can provide the reader with the most relevant information for their situation. 

An interactive map uses reader's current location to show information about the income mobility in their area. Courtesy The New York Times.

When allowing AI to report news in this algorithmic system, however, there is a potential lack of context provided to the reader. AI cannot gather historical background, identify cultural implications, or relate to the reader in the same way as a human journalist. Journalism has a long history of adapting to technology and the adaptation to AI technology is no exception. “Technology does not become, replace or stand for journalism,” rather, journalists use technology as a tool to produce truthful reporting (Zelizer 3). Journalistic jobs will not be impacted by the addition of AI to write news articles if news organizations give technology the space to be used as a tool (Zelizer 5). In addition to writing articles with AI, news and media organizations are using AI to work as news bots on social media.

            A news bot is a piece of artificial intelligence that is designed to gather articles from news sites and distribute them. "Bots often appear as productive social agents that populate chat apps and social network platforms such as Twitter, Facebook, and Reddit," (Diakopoulos 147). News bots are beneficial for individual news outlets to generate interest in their content through social media. McBride and Rosenstein argue that “to maximize the effectiveness of published content, journalists need to understand the networked nature of audiences” (142). Followers and readers of the news are interconnected and have the technological tools to give feedback and start Twitter conversations about the news pieces they read. News bots have been designed as tools that journalists can use to facilitate conversation about the news and share links to their articles on social media. The New York Times and other news organizations have used news bots to redistribute their own articles on Twitter to target different target audiences (Diakopoulos 150). 

            News bots have a more powerful role, however, when they are used by news aggregators. News aggregation is "the practice of redistributing news content from different established news outlets on a single website” (Lee 5). These aggregation sites include Google News, Yahoo News and the Huffington Post, among others. In traditional newsrooms, a committee of human editors serve as the gatekeepers of information by deciding what to include in the print newspaper and in which order. In broadcast journalism, the TV station or the news anchor are the gatekeepers of news by deciding what to air for their audience. On the internet, however, AI acts as an invisible gatekeeper to our information. News bots and AI often determine which information we see on social media and on news sharing websites. Information is now being categorized based on personal preference and click-ability, rather than the importance of that news on a larger scale. 

Google News is an example of an invisible gatekeeper of information, recommending articles based on the source, recency of the article and number of related articles. Courtesy of Google News.

            The digital age has made information more available than ever before, so many news readers are coping by allowing bots to sort through the available news and organize headlines for them. The AI uses different "news cues" to determine which information is important for news consumers to see (Sundar 3). Google News considers three news cues: name of the source, recency of the article and the number of related articles (Sundar 3). These three characteristics of the article are visible to the viewer when browsing Google News. The intention of using these cues is to convince news readers that the bot has provided the most important news at that moment. The responsibility now falls on media consumers to determine whether these cues are the best determinants for their own news consumption. Are the stories with the highest credibility always going to be the most relevant? Is newsworthiness always equated with recency? Does the number of related articles directly relate to the importance of the story? These questions do not have definitive answers, but many news readers are allowing bots to answer these questions for them about the news they consume. News aggregator sites give news consumers the impression that they are finding the information themselves, which gives people greater trust in the news they read (Sundar 4). For news readers, searching through the group of news articles curated by a news bot can feel like they discovered the information on their own. A news reader could not even consider engaging in articles outside of the ones harvested by AI for their own efficiency.

            News aggregators like Google News also have the possibility of only recommending articles that are niche for a reader’s interests or political leaning. For example, an article about a research organization announcing a new study about mileage on car tires could be tailored to your particular car and the number of miles you drive (Chace). Consumer’s data is already being collected every day for companies to curate their advertisements to consumer’s needs.

Now, reader’s personal data could be embedded in the very news articles that they trust. Just as the New York Times published in article in 2015 that contains an interactive map that uses the reader’s current location to curate the numbers in the article to the reader’s location, news aggregators have begun to use AI to recommend articles that relate to specific data about consumers. 

            AI has helped journalists write stories, allowed news aggregators to provide personalized recommendations and given news bots on social media the power to spread misinformation. The responsibility falls on journalists to harness AI and bots to communicate accurate and truthful information on social media and in their writing. During the COVID-19 pandemic, the effects of a misinformed public due to false information being spread through news bots has been evident. Since journalists and citizens both have access to the technology of bots, there is an even playing field on social media for sharing information. This can have positive effects for journalists, such as engaging and connecting with their readers through online discussion. There are also potential dangers for journalists, who might worry that AI could become the prominent voice of news, rather than the human journalists themselves. 

Journalists can view journalism as a practice that gives technology purpose, shape, perspective, meaning and significance, according to Zelizer (8). Rather than seeing technology as a force that is going to take over journalism, Zelizer argues that journalists can be thoughtful about what place they give technology in their news practices (8). Twenty years ago, the internet was seen as the biggest threat to journalism because it broke down the existing business model of ad-revenue. Journalists adapted, however, and have established online news as the primary source of information for the majority of Americans. News organizations also adapted to subscription-based business models to survive when Google and Facebook took most of the ad revenue from print newspapers. Many news organizations face another shift in technological advancement today with the rise of AI. Journalists are working to harness AI’s power to their advantage in storytelling. “If journalism is to thrive productively past this technological revolution and into the next, we need to do better in sustaining a fuller understanding of what journalism is, regardless of its technological bent, and why it matters” (Zeilzer). 

Works Cited

Allyn, Bobby. “Researchers: Nearly Half of Accounts Tweeting About Coronavirus Are Likely Bots.” NPR. 20 May 2020, https://www.npr.org/sections/coronavirus-live-updates/2020/05/20/859814085/researchers-nearly-half-of-accounts-tweeting-about-coronavirus-are-likely-bots. Accessed 14 April 2021. 

 

Chase, Calum. “The Impact of AI on Journalism.” Forbes. 24 Aug. 2020, https://www.forbes.com/sites/calumchace/2020/08/24/the-impact-of-ai-on-journalism/?sh=3bf4f07a2c46. Accessed 1 April 2021.

 

Diakopoulos, Nicholas. Automating the News : How Algorithms Are Rewriting the Media, Harvard University Press, 2019. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/lmu/detail.action?docID=5761259.

 

Lecompte, Celeste. “Automation in the Newsroom.” Nieman Reports. 1 Sept. 2015, https://niemanreports.org/articles/automation-in-the-newsroom/. Accessed 2 April 2021.

 

Lee, Angela M. and Chyi, Hsiang Iris, “The Rise of Online News Aggregators: Consumption and Competition,” International Journal on Media Management. 2015, https://www.tandfonline.com/doi/abs/10.1080/14241277.2014.997383

 

Levenson, Eric. “L.A. Times Journalist Explains How a Bot Wrote His Earthquake Story for Him.” The Atlantic. 17 March 2014, https://www.theatlantic.com/technology/archive/2014/03/earthquake-bot-los-angeles-times/359261/. Accessed 14 April 2021.

 

McBride, Kelly and Rosenstiel, Tom. The New Ethics of Journalism. Thousand Oaks, CQ Press, 2014. 

 

Pressman, Laura. “The Automated Future of Journalism.” Automated Insights. 28 Feb. 2017, https://automatedinsights.com/blog/the-automated-future-of-journalism/. Accessed 14 April 2021.

 

Roberts, Siobhan. “Who’s a Bot? Who’s Not?” The New York Times. 16 June 2020, https://www.nytimes.com/2020/06/16/science/social-media-bots-kazemi.html. Accessed 14 April 2021.

 

Sundar, S.Shyam, et al. “News Cues: Do Indicators of Newsworthiness by Newsbots Affect Our Perception of News Stories?” Conference Papers -- International Communication Association, May 2005, pp. 1–34. EBSCOhost, search.ebscohost.com/login.aspx?direct=true&db=ufh&AN=18655521&site=eds-live&scope=site.

Zelizer, Barbie. “Why Journalism Is About More Than Digital Technology.” Routledge Taylor & Francis Group, vol. 7, no. 3, 2019, pp. 2-8.