Home Computer Science
Fake news, Al, and digital platforms
While media organizations, both newspapers and broadcasters, have continued to develop and advance their Al systems to resonate with their journalistic standards, although limited to a few mega companies, the use of Al in social media platforms has shown a very different direction. As discussed, automated journalism in newspapers and broadcasters works for fact-based stories for which clean, structured, and reliable data are available (Graefe, 2016). However, Al use in digital platforms in tandem with journalism is much different from traditional journalism, as they create fake news with no fact-based stories.
There has been no doubt that the rise of social media has transformed how people consume news and information. Unlike traditional media, including newspaper and broadcasting, social media platforms expedite the fast circulation of news as users share news and information at the click of a mouse. Due to their unique functions and capacities, social media platforms play a key role during natural disasters by disseminating news and warning signals as quickly as possible. The general public also share news, while expediting the circulation of any information through online and eventually offline friends. For example, Facebook comes with a share function, and “users can either share an article on their own page, or post it on a friend’s page, or send it through private messaging either to a particular friend or to a group” (Tandoc et al., 2019, 674).
In contrast to several positive possibilities of social media platforms as alternative media, they have instead brought about several negative impacts to contemporary society. Again, the rapid circulation of news through digital platforms has “led to widespread and effective forms of media manipulation” (Pangrazio, 2018, 8). Among these, fake news has deeply influenced not only people’s daily lives but also the news ecosystem itself. Fake news is not new, but has rapidly grown since the mid-2010s. As Table 7.1 clearly shows, newspaper articles on fake news around the globe have increased, from 2,010 cases in 2015 to almost 60,000 cases in 2018.2 This data, although limited to newspaper articles that can be accessed via a Lexis Uni database, certainly proves the recent surge of fake news in our society.
What I emphasize here is the critical nexus of fake news and Al, as Al plays a key role in producing and circulating news. Due to their enormous influences, news organizations are right to consider the rise of fake news a serious issue. “It not only has the potential to mislead large numbers of people, but also to draw audience views and engagement away from real news produced by legitimate news organizations” (Tandoc et al., 2019,673). Of course, Al technology is also able to function to block fake news on social media platforms, although this function is currently limited.
Table 7.1 Number of newspaper articles mentioning fake news, 2011—2019 Source: Lexis Uni Database.
Increasing fake news in the Global North
There are several precedent cases of fake news; however, people have mainly begun to pay attention to fake news because of politics. The role of Facebook as the predominant social media platform that spreads fake news became clearer on a large scale during the 2016 U.S. presidential elections. Back then, Facebook played a role in spreading false and misleading information, known as “fake news,” from Russia and from inside the U.S. about candidates. Ever since then, the dissemination of fake news has continued to rise. As the Associated Press (AP, 2019) reports, Avaaz, an advocacy group that tracks misinformation, says it has found an increase in fake political news shared on Facebook ahead of the 2020 presidential elections. Avaaz found that, collectively, fake stories were posted more than 2.3 million times and had an estimated 158.9 million views, along with 8.9 million likes, comments, and shares. Most of the false news sources were individual users or non-official political pages. Avaaz said in the report that the findings are the “tip of the iceberg of disinformation” ahead of the 2020 elections (Associated Press, 2019).
The blame mainly goes to social media platforms, of course. As is well documented (Pangrazio, 2018, 12-13),
The goal of many platforms to personalize, customize, and tailor user experience, means individuals become aligned with social groups that
Al journalism, social media, and fake news 125 can lead to a kind of group think approach to news and information. This also means factually incorrect information, or fake news, can be rapidly spread through social groups and networks which share the same or similar beliefs and values.
Fake news can be produced because of the combination of the users who want to make money in a broad sense (e.g., influence and power as well as capital) and consumers who share the information and news as they believe them to be true through social media platforms. Social media platforms in this sense mediate two different users to garner monetary profits and network power.
Due to several occasions of fake news related to the presidential election and privacy invasion, Zuckerberg had to testify before the U.S. Senate in April 2018. In his opening remarks Zuckerberg said that Facebook did not do enough to prevent its tools from being used for harm—citing fake news, interference in foreign elections, hate speech, and data privacy (CBC News, 2018). Zuckerberg promised to deploy Al to help solve some of the company’s biggest problems, by policing hate speech, fake news, and cyberbullying, although it has been an effort that has seen limited success so far (Knight, 2019).
Facebook has consequently advanced its fact-checking system to curb fake news on its own platform. In 2019, Facebook (2019) itself announced it removed 3.39 billion fake accounts from October 2018 to March 2019. Facebook said nearly all of the fake accounts were caught by Al and more human monitoring. They also attributed the skyrocketing number to automated attacks by bad actors who attempt to create large volumes of accounts at one time (Romo and Held, 2019).
Whether creating or blocking fake news, the complexity of fake news is intrinsically intertwined with the logic of the Al algorithms behind social networks, including Facebook.
The mechanisms of personalization and tailored content play a special role in this context, possibly creating filter bubbles—in which intellectual isolation is caused by digital algorithms that selectively assume what kind of information the user would like to see—and echo chambers, in which closed media systems reinforce beliefs.
(Borges and Cambarato, 2019, 610)
The issue of fake news is further “exacerbated as social media sites, along with other online platforms, are used to deliver and generate political news and information to ideologically segregated audiences through the use of sophisticated geotagging and microsegmentation strategies” (Brummette et al., 2018,498).
More specifically, Facebook’s own news feed was originally born in September 2006, promising to provide a personalized list of news stories throughout the day. In the early stage of the service, news feed ranking was turning knobs: turn up photos a little bit, turn down platform stories a little bit. Facebook advanced from “turning knobs” to EdgeRank, the algorithm that a) determined which of the thousands of stories (or “edges” as Facebook called them) qualified to show up in a user’s news feed and b) ranked them for display purposes. EdgeRank had three primary pieces: 1) affinity (i.e., how close is the relationship between the user and the content/ source?); 2) weight (i.e., what type of action was taken on the content?); and 3) decay (i.e., how recent/current is the content?). EdgeRank made it possible for Facebook to give users a more personalized news feed (McGee, 2013). For example, users who played many games on Facebook could see more game-related content in their news feed. But now that job is a lot more complicated than ever. In fact, in 2013, Facebook changed to a machinelearning algorithm that considers more than 100,000 weight factors when producing news feeds (McGee, 2013). In 2018, Facebook, one more time, overhauled its long-term plan, which was “prompted largely by feedback that posts from friends and family having been increasingly drowned out by content from brands and publishers, deterring people from interacting more with people they care about” (Chaykowski, 2018). Facebook equipped with Al has continued to develop new functions to attract more users, which means that the personalization of items has been intensified through a delicate Al system.
Fake news and regulation in the Global South
Fake news is not only limited to Western countries; it is also common in non-Western countries. As social media platforms themselves are global, fake news is occurring everywhere. Several Asian countries like Korea and Singapore have experienced the rise of fake news and planned to regulate fake news through various measures, which are sometimes controversial because these regulatory mechanisms may hurt democracy as well. Korea, for example, has experienced a surge of fake news on social media platforms since the mid-2010s. While stories vary, fake news related to politics has continued to grow, which demands that the Korean government take action. The Korean government planned to crackdown on fake news, calling it
a destroyer of democracy. Speaking at a cabinet meeting, former Prime Minister Lee Nak-yon said that fake news had spread so widely in Korea that it was stymying not only citizens’ privacy but also the country’s national security and foreign policies, including its relations with North Korea. Lee continued to say, Fake news is a public enemy hiding behind the cover of free speech,
and therefore, we can no longer turn a blind eye to it.
(Chae, S.H., 2018)
Mr. Lee told the Korea Communications Commission (KCC), a government regulatory agency, to act on online media sources that serve as avenues for fake news.
Consequently, KCC called for the creation of more institutions dedicated to checking whether or not certain information is based on fact, in a bid to stop the spread of fake news in the nation. Han Sang-hyuk, who took the office as KCC chairman in September 2019, made the following remarks during his first luncheon with journalists, “Both ruling and opposition parties, as well as members of the public, are all aware of the problem of fake news; thus it is surely necessary to come up with proper countermeasures.” He said, “I believe that boosting fact-checking institutions trusted by the public could be one measure; however, it would be improper to establish such an institution within the KCC as it could provoke concerns over government infringement of freedom of speech” (Jun, J.H., 2019c).
However, the crackdown on fake news is very controversial due to its potential to curb the freedom of speech. Lee Hyo-seong, the former KCC chairperson, resigned partially because of the disputes surrounding his role regarding the handling of fake news. The government demanded a very strong policy measure to curb fake news; however, Chairperson Lee of the KCC emphasized voluntary regulative measures. Lee, who was a critical media scholar before taking over the position, continued to emphasize that “outright fake news certainly negatively influences the people; however, it should be controlled by voluntary measures developed by civic activities instead of regulatory laws” (Kim, J.H., 2019).
Meanwhile, Singapore has enacted a law to control fake news. In November 2019, Singapore passed a controversial anti-fake news law that gives authorities sweeping powers to police online platforms and even private chat groups. The government can order social media platforms to remove what it deems to be false statements that are against the public interest and to post corrections. The Singaporean government has emphasized that the law would not be used to target opinions, but only falsehoods that could prove damaging. “Free speech should not be affected by this bill,” Law minister K. Shanmugam told parliament, adding that the law is aimed at tackling “falsehoods, bots, trolls, and fake accounts” (Wong, 2019). However, critics claim that it is a serious threat to civil liberties.
Interestingly, based on the new law, the Singaporean government asked Facebook to issue a correction in November 2019, which was actually implemented. Facebook has added a correction notice to a post that Singapore’s government said contained false information. It is the first time Facebook has issued such a notice under the city-state’s fake news law (Butcher, 2019). Singapore claimed the post, by fringe news site States Times Review (STR), contained scurrilous accusations. The note issued by the social media giant said it is “legally required to tell you that the Singapore government says this post has false information.” Facebook’s addition was embedded at the bottom of the original post, which was not altered. It was only visible to social media users in Singapore (BBC News, 2019).
As discussed thus far, there are mainly two types of Al use in journalism. One is adopted by news organizations themselves, and the other is used by digital platforms—in particular, social media platforms. As the producers and distributors of news and information, these media organizations and social media platforms have certainly adopted and utilized Al in writing and distributing news and information. Whether talking about media organizations or social media platforms, what we have to focus on is whether Al has fundamentally transformed the nature of journalism and what the consequences are. Al-equipped social media platforms function as both the creators and disseminators of fake news, including political news. Fake news in these non-Western countries on social media platforms has been increasing, as with Western regions. These non-Western countries do not benefit from the growth of Al in journalism but heavily experience the increasing role of social media platforms in spreading fake news that disturb contemporary society and democracy.