Is Robot journalism the future of the news?
A study of the role conceptions of journalists and the implementation of automated journalism into the newsrooms
Charikleia Kotsai Student ID: 12846821
Master’s Thesis
Graduate School of Communication Master’s programme Communication Science
Erasmus Mundus
Journalism, Media & Globalization Supervisor: Sanne Kruikemeier
29-05-2020 Word Count: 6,409
Abstract
The unprecedented amount of data leads newsrooms to implement automated journalism in order to manage the huge amount of information that circulates on the Internet. This study extends the literature on the subject by using Grounded Theory to investigate to what extent does algorithms affect journalists and their role
conceptions. This study brings clarity to a field in flux, by exploring the attitudes of human journalists towards adopting automation in the newsrooms and analyzing their views about the changes in their traditional role conceptions. Through in-depth interviews with professional journalists, this study reveals that automation adoption increases concerns about the rise in the autonomous power and the uncertainty for the future of the journalistic field.
Keywords: automated journalism, role conceptions, newsrooms, bias, algorithms, news-writing programs, technology and automated-writing software.
Introduction
In this era of automation and “big data”, when information disseminates in unprecedented ways and algorithms can create stories, newsrooms worldwide are starting to embrace and adopt automation (Wu et al., 2019). Automated journalism has made its appearance in the last years in the journalism industry. Scholars have conceptualized automated journalism as algorithms that create news stories with limited to no human intervention (Carlson, 2015). The development of artificial intelligence (AI) has made its introduction and integration across many industries the last decade. The newsrooms are adapting to the challenges of AI and the professional human journalists are trying to adjust on the incorporation of data and algorithms in journalism; automated journalism (Lewis et al., 2019).
Automated journalism is based on algorithms and can produce more stories than a human journalist can do. “Automation anxiety” (Linden, 2017, 24) is one of the many feelings that journalists are dealing with since robotic journalism made its first appearance. The Artificial Intelligence integration in society is changing journalism by making it more co-effective and cheaper (Aljazairi, 2016). Carl-Gustav Linden (2017) estimated that having algorithms producing stories, that contain less errors, put journalists in an uncomfortable position, since journalism employment has dropped dramatically the past years. According to Linden (2017), in 2013, in the UK, around 70,000 people were employed as “journalists newspaper and periodical editors” (p. 124) but in 2015, these numbers decreased by 6,000 people before the financial crisis started.
Scholars have found that newsrooms are not automated yet, but might become more automated once it is deemed to be of perceived relevance to journalists
(Carslon, p. 412). According to Carlson (2015) due to the rise of technological developments, it raises issues of “the future of journalistic labor, the rigid conformity of the news compositional forms, and the normative foundation of journalistic
authority” (p. 417). Montal and Reich (2017) proposed two different levels of algorithmic contribution in journalistic content. The first level is algorithmic content generation created with no human journalist/editor involvement and the second level is integrative content generation, which indicates that the production of a textual or visual story needs the collaboration of one or two human journalists and a generative algorithm.
All of these developments and the rise of robotic journalism have created implications on human journalists and raises concerns about its reliability and
of automated journalism affects journalists and their role conception will offer insights into the challenges that the traditional perceptions of journalism face. An artificial intelligent algorithm can replace a human journalist in some tasks that thinking and decision making are required (Montal & Reich, 2017).
In sum, the purpose of this study is to provide an initial framework for
understanding the transformation of the newsroom and the quality of journalism. The study will examine the robotic journalism functions and how this affects human journalists from the perspective of journalists. The main research question will be: To what extent does automated journalism affect journalists and to what extent does it affect their role conceptions?
A series of interviews with journalists were conducted with the aim to
understand their perspectives about the role of automated journalism. Building on the theory of journalistic role conception, experts have found that automated writing software has changed the ways that communication is mediated between people and the flow of information has been transformed among media producers and consumers (Lewis et al., 2019). This could indicate that the new way of communication might affect the field in a positive or negative way based on the approach of the newsroom towards automated journalism.
Automation in the Newsrooms
The introduction of automated journalism and digitalization raises a question about the quality of journalism. Computational journalism faces ethical challenges regarding the error with underlying data or automation software, credibility and accuracy of the stories that are being created by robots (Kent, 2015). Young and
Hermida (2014), defined computational journalism as a combination of systems, algorithms, social science and math, creating news.
The implementation of Artificial Intelligence in the newsrooms can create mixed feelings like doubts and a sense of fear that can rule journalists (Lemelshtrich Latar, 2015). Based on Lemelshtrich Latar (2015), the newsrooms are undergoing dramatic changes, such as news way of functioning; adopting automation in the newsrooms is changing the way that traditional journalists and the news outlets work. Since it is important for the employers to know that the journalism ethics that exist now, separate facts from opinion but automation journalism ethics have not yet been written. It is important to stress out that the publishers should make it transparent, which article was written by an algorithm and which by a human journalist. The differences between them are significant, and the readers might want to know who writes what. Media organizations are expected to be efficient. Once the newsrooms fully permit the use of robotic journalism with the aim to create more content and write news stories in a short period of time, then concerns like liability and the ethical procedure of journalism are expected to arise. Artificial Intelligence algorithms will be employed to conduct an automatic analysis of the consumer’s segmentation and tagging them, with the aim to analyze the “consumer engagement during the content consumption for a layout of an automatic newsroom” (Lemelshtrich Latar, 2015, p. 74).
One of the many reasons that some newsrooms prefer to implement algorithms is that they can generate a lot of content in a short period of time. Although it is difficult to identify and isolate bias Ombelet, Kuczerawy, and Valcke (2015), criticized that algorithms are not neutral, meaning that software is biased, while Whittaker (2019), supports that journalists and the media organizations are also
operating with some kind of bias. Newsrooms function differently in Europe and differently in the United States. However, there is a difference between the traditional European journalists and the traditional U.S Journalists. It is estimated that Europe has a closer association between journalistic partisanship and that of their news organizations (Patterson & Donsbach, 1996).
Another one is that algorithms can be adjusted depending on the reader’s preferences and profile, while it is predicted that in the future automated journalism will succeed in creating multiple versions of a news story with the aim to match the individual’s profile and tastes (Ombelet, et al., 2015). Once the newsrooms fully allow robot journalism to create content and write news, then concerns like liability and the shrinking of human reporters will arise (Ombelet, et al., 2015). Media organizations are expected to be efficient. Artificial Intelligence algorithms will be employed to conduct an automatic analysis the consumer’s segmentation and tagging them, with the aim to analyze the “consumer engagement during the content
consumption for a layout of an automatic newsroom” (Lemelshtrich Latar, 2015, p.74) Adjusting automation into the newsrooms can result the perception of it in two different perspectives by journalists and editors. The first perspective is that
automation will be the main factor and can threaten the job of a journalist and the second one is that automation can be viewed as complementary, so humans and machines can work together (Wu, et al., 2019).
The perspective of a Human Journalist towards software
The algorithmic process can be described by three waves. The first wave is an ‘Algorithm wave’, which is doing computational tasks that are not complex and it structures and analyzes data. The second wave is called ‘Augmentation wave’, which
is contributing to the change of how many job tasks are being performed. The third wave is the ‘Autonomy wave”, which introduced AI and robots (Evans-Pughe, 2017). The ‘Autonomy wave’ (Evans-Pughe, 2017) is the one that has started to take place to the media organization, maybe is not very obvious yet because the automated
journalism is growing slowly and to some countries is not known yet, but it is. The lack of credibility in the medium, source and message seems to face robots since as Wolker and Powell (2018), stated the credibility of the message worries content, the credibility of the content worries the writers and the credibility of the medium concerns the way in which the message is conveyed. However,
sometimes the readers are not capable of confirming every message they receive, and that is why a credible journalist is necessary in order the individual to receive
“accurately mediate reality” (Powell & Wolker, 2018, p.3). Knowing that a human-journalist writes the article raises credibility and readability ratings. Some human-journalists may view the rise of algorithmic processes as a threat, thinking that it will eliminate jobs but another perception is that automated journalism might help journalists by finding interesting stories and provide them with a draft that they will work on after (Graefe & Halm, 2016). Based on the study of Thurman, Dorr, and Kunert (2017), professional journalists disagree about the function of software. Some of them claimed that it lacks human angle and traditional reporting. The limitations that automated journalism has, makes the human journalists view it as less creative compared with traditional journalism. Others were quite satisfied by mentioning that algorithms can be complementary to their work and it has potential to increase the speed, thus it can expand the volume of coverage.
Human journalists observe that robots are going to constitute a strong
turn them into readable stories without human involvement. If robot journalism can be programmed in a correct way, than it will not miss facts and if programmed
objectively then are free of bias (Lemelshtrich Latar, 2015). The cultural attitudes of journalists regarding artificial intelligence can influence the conceptualization of automated journalism. Automated journalism might replace journalists or justify not hiring one (Lewis et al., 2019). The optimistic view of automated journalism is that the algorithmic process can provide to the journalist a first draft including the basic facts that they will be able to enrich and verify (Lemelshtrich Latar, 2015). So, the involvement of a human journalist is very important because the analysis that they will provide will be deeper and will put a higher value on it (Graefe et al., 2018). The pessimistic view of the algorithmic process is that it can be a threat to human
journalists' working lifestyle (Lemelshtrich Latar, 2015). The progress of the algorithms can be interpreted as a scary process that makes journalists doubt about their professional future or as a helpful tool that makes the process of the news faster. Lemelshtrich Latar (2015), supports that when it comes to the adoption of algorithmic processes, human journalists have to understand the limitations that robotic
journalism has and take advantage of it. They can use the Artificial Intelligence algorithms and obtain more power by “ learning to adapt the nature of their work to the new social physics Big Data era” and also adds that some journalists “see robotic journalism as an opportunity to make journalism more human” (Lemelshtrich Latar, 2015, p. 76). An algorithm lacks the ability of critical thinking, which leads to not being able to write a story the same way that a human will, because it cannot understand complex issues, cultural context and be empathic when it is needed. However, AI algorithms can enhance innovation but it cannot create one, for this human integration is needed. (Lemelshtrich Latar, 2015, p. 77). Within cyber ethics,
other fields of ethics like “new media ethics” (Debatin, 2010) or “digital media ethics” (Ess, 2009) exists and it is important to understand the ethical discourse in Robot Journalism. The values of journalism such as objectivity, neutrality,
transparency, are changing. Algorithmic journalism is being criticized because it has a variety of factors and different kinds of responsibilities. Data is raising ethical issues about its origins and its limitation on items that put value on the data set, while journalists are worried about personal privacy rights and the coding part since it reflects the journalistic values (Dorr and Hollnbuchner, 2017). So while we assume that the implementation of AI in the newsrooms is transforming journalism by introducing something new, this study moves beyond current literature by investigating how automated journalism affects journalists based on these considerations, the following research questions are proposed.
RQ1: How do journalists within newsrooms view automated journalism?
RQ2: How do journalists see journalism changing due to the algorithmic process?
The role conceptions of journalists and algorithmic journalism Numerous studies have been conducted with the aim to analyze how
journalists around the world view their professional tasks. Journalists value their role of conception since they tend to identify their identity in the journalistic profession. (van Dalen, de Vreese, & Albæk, 2012). Shoemaker and Reese (1996) claim that the journalistic roles can influence the news content directly since “these determine what the communicator thinks is worth transmitting to his or her audience and how the story should be developed” (van Dalen, de Vreese, & Albæk, 2012, p. 905).
Coleman and colleagues (2018), separate the role conceptions of journalism into four categories. The first one is the interpretive or investigative function that involves analyzing complex topics and digging deep into governmental issues, the second one is the role of a disseminator which is defined by collecting fast
information and disseminating it to a large audience and offering entertainment when it is needed. The third one is the adversarial role, which actually means being critical when it comes to actions executed by public officials and businesses. The last one is the role of populist mobilizer, who is setting the political agenda and trying to produce interests such as intellectual or cultural.
However, the study of Skovsgaard and colleagues (2013), points to the significance of objectivity on the role conceptions, since it is being used to persuade the audience about the credibility and validity of the news story. They also mention two distinctions in order to show the different role conceptions of journalists. The first one is the “aggregation of preferences”, which is linked to the idea that journalists are collecting necessary information about citizens, who are living in democratic societies and casting their votes. The second one is an approach of being a passive journalist, which believes that events can be perceived and reported as they are and he or she does not want to invent, and an active function, which means that the journalists take action in the news production process.
The link between journalists and their conceptions can be influenced by the cultural attitudes that they already have. A survey conducted by van Dalen, de Vreese, and Albæk, (2012), showed that Spanish political journalists were different from their colleagues in Denmark, Germany and the UK. However, the only similarities that they might have are in terms of reporting style and the frames they use. Journalists coming from different cultures follow different values linked with their profession
and that leads to a different journalistic function. A research from 2017 found that the role of interpretive was still considered as one of the most important roles in
journalism, while the importance of disseminator marked a decline and the role of populist mobilizer has increased. The study concludes that the role conceptions of journalists are being influenced by “sourcing practices and idea generation routines for citizen contributors” (Chung, & Nah, 2013, p. 90).
The introduction of Artificial Intelligence into journalism has changed the communication dynamics, and made it less familiar. The increase of automated journalism has grabbed attention towards the ways that algorithms are changing relationships and the flow of information among media producers and consumers, and also the way that communication is mediated between individuals. (Lewis, Guzman, & Schmidt, 2019, p. 412). Based on the perception of social-interaction journalists are explaining their actions through communication, interaction with sources, audiences and media organizations (Ojala & Pöyhtäri, 2018). The role conception of a journalist depends on the environment that surrounds him or her, since the expectations that the environment has, it might be different from the expectations that the journalist has. Data-driven journalism will constitute a strong competition for traditional journalists (Lemelshtrich Latar, 2015) and will lead both of them to find a way to negotiate and combine the roles (Ojala & Pöyhtäri, 2018). As the machines are becoming smarter, human journalists feel the urge to protect their values, autonomy and the journalistic field. Competing and pointing that software is prone to errors is one of the many reactions that journalists have (Wu et al., 2019). Role conceptions are influenced and formed by the professional experiences that journalists have and learn during their job. Algorithms are involved on the production of the articles by generating data and writing stories that might not be reflected as credible stories (Wu et al., 2019) and that
might have an influence on the roles of journalists (Dalen et al., 2012), since they will have to deal with the lack of credibility and transparency in the future. This study assumes that automated journalism has a negative impact on the role conceptions of journalists and put under a dangerous position the role of disseminator. Based on these assumptions the following research question is being suggested to analyze and discuss further.
RQ3: How do journalists see the role of automated process in relationship to the role conceptions of journalists?
Methodology
This qualitative study analyzes journalist’s perceptions and opinions about automated journalism. Semi-structured and open-ended interviews were conducted with ten journalists who were aware about the automated journalism in the
newsrooms. My primary goal was to have fifteen participants but because of the Covid-19 crisis it was very hard to find journalists who would be willing to talk and also automated journalism is not that well known yet to some journalists, so that made it more difficult to find journalists who would like to participate. Individual
interviews were chosen because of the ability to go deeper in the issue matter and have more detailed answers when it comes to personal experiences that some of them had with automated journalism.
Sample Selection
Purposive sampling was used to recruit ten individuals who were selected based on their occupation and having knowledge about the topic. The sample is mixed, which means that the journalists are not picked from one specific country. In this study, journalists from: Greece, the Netherlands, Belgium, Brazil, Albania, Ukraine, the United States, and Nigeria participated. Most of them are working
abroad and not in their home countries. The age range is from 20 to 60 year old, working for broadcast, print, podcast and online media (see Table 1).
Table 1.
Characteristics of the participants
Code Age range Occupation Media type
A 20-30 Journalist/Researcher Online
B 20-30 Financial Journalist Online
C 20-30 Journalist Broadcast
D 20-30 Radio Journalist Podcast
E 20-30 Freelance Journalist Online
F 30-40 Freelance Journalist Online
G 30-40 Journalist Online
H 30-40 Freelance Journalist Online
I 50-60 Journalist/Journalism Trainer Online
J 50-60 Journalist/Media Educator Print
The Skype interviews were recorded using my phone and were all held in the third, the fourth week of April 2020 and the first week of May 2020. The interviews lasted between 18 to 35 minutes.
I chose a semi-structured open-ended interview, because I did not want to have a specific order for my questions but rather, to have the flexibility to let the conversation flow and pick the question that suits the right moment. I conducted the interviews via Skype with journalists. Using qualitative research based on interviews benefits my study in terms of getting new information since the topic is quite new. Scholars have studied the threats of automated journalism or the implications of it on journalism but they have not studied in which way it affects the individual journalist. Ground theory was used with the aim to support this research.
Within this topic, the aim was to explore journalist’s views and opinions on the issue, and to understand if they perceive the introduction of automated journalism as a threat or a positive evolution of journalism. Therefore, the participants were asked why the newsrooms were using automated journalism and how they are feeling about the integration of Artificial Intelligence in journalism. To ensure internal validity, quotes of the interviews were provided in the result section.
Analysis
Grounded theory was used as a methodological framework in order to support the analysis. The theory is a set of systematic inductive methods, which means
specific guidelines, strategies data collection and analysis (Charmaz, 2009). Verbatim transcriptions were conducted to the interviews and were analyzed as axial coding, which is a qualitative technique that develops linkages between data (Allen, 2017)
suitable for this study, because the participants raised common topics and justify their opinion in a similar way. So, the linkages between their opinions were easier to make and analyze. During the process of the axial coding, different categories were created based on the common issues that the participants were discussing during the
interview. The aim of the categories was to make the connections of the common opinions in order to be able to analyze and summarize that the participant journalists had in order to provide clarification what was important to mention in the current study and give answers to the research questions. In the beginning of the analysis more categories, like ‘negative impacts on the reader’s’ and “no impact on the reader’s”, were created in order to analyze the perspective of the journalist regarding the impacts that the articles written by an algorithm, if any, will have to the reader. But during the process of the analysis, these codes were not necessary anymore, since they could be analyzed and summarized on the positive views or the negative views. The deductive codes were based on the four categories: 1.Negative views, 2.Positive views, 3.Solutions, and 4.Bias. The issues that the participants raised and after they justified created the inductive codes.
Findings
The power of automation in the newsrooms
RQ1 asked about how newsrooms perceive automation. The journalists that were interviewed defined automated journalism as algorithms that combine facts in a syntactical way that people can understand but will be lack of explaining to people why something matters. Although, the automated journalist can write an article that contains all the facts in a readable way, it will still not be able to explain further by giving meaning and opinion about the topic or explaining how this issue affects the
reader. “Good writing contains journalists’ judgments and personal feelings” (Code J) and that’s why articles written by an automated journalist will not be able to reach the level of a piece written by a human journalist. All the participants replied that the newsrooms benefit from the automation, since they see value in it. As participant B said: “It is quicker, I’m sure you can save a lot of time and it’s cheaper as well, because you don’t need to have to pay journalists to do that. [….] I do see that there is definitely value if you have content automated based on a database for example, because the computer can make those connections much better than a human being, so there you definitely see value.” (Code B). News organizations have found
themselves in an advantageous position and they understand the benefits of having automated journalism, it can generate more content than a human can, with no spelling mistakes and as the participants said it is affordable and quick. Newsrooms seek efficiency, which lies in AI algorithms by sending fast content directly to the readers. The participant C states “I suppose that nobody will be happy about this, because jobs will be lost but I think it’s a bit…sad that robots will be able to create content because this was always something human”. Implementing automation in the newsrooms raises questions about the differences between an article written by a human journalist and an article written by an algorithm. One of the participants said that“ I would be very suspicious and actually a bit angry, because I expected to read a piece that a professional journalist has done. [...] I think I would be very frustrated and disappointed as a citizen. A professional human journalist is the one that gives meaning to the information” (Code A). Moreover, the same journalist refused to call the algorithms as with the name of an automated journalist instead insisted calling it a machine “[…] since we do not have an author in the article, I would not call this journalism”. The journalists were concerned that no authorship can bring problems,
since if a fact is written wrong in the article, people will not be able to put the blame on someone. “Whoever publishes news there should be someone that is accountable for the way that it is portrayed and not just say that this is automated content. And if there is something not properly reported you couldn’t accuse anybody for it because there is no a human being behind it and I think it is a major issue” (Code B). Articles by algorithms are lacking transparency, accountability and narrative aspects. Most of the participants support that a robot lacks identity since it does not have an author, so people will not be able to understand who controls the process and if any mistake occurs they will not have anyone to blame. Some of the journalists claimed that an article created by an algorithm would not affect the readers, since they will not be able to understand it; the other participants said that it will have an impact on them and the reader will feel “disappointed”, “betrayed”, “cheated” and “angry”. These automated writing technologies, if programmed correctly, seldom are missing facts or making spelling mistakes and actually, it can be bias free, however, the newsrooms and the software engineers are the key for that to succeed. The majority of the participants supported that algorithms are not biased but if the input is biased then probably will be biased. Algorithms cannot contain any bias, but the way that the newsrooms use it can change that. One of the participants, who worked with an automated journalist said, “the people that are creating the algorithms are mostly predominantly male, white engineers, who have their own biases”(Code B). Also, the same journalist gave an example of an article from the New York Times, which was addressing the topic about how police forces try to implement AI in their searches. This article indicated that the data were less accurate when it’s about black people than when it is about white people. The person, who decides where the algorithm should look for in order to compile facts and create the story, is the one who creates
the bias. “People tend to think that technology is flawless and technology is not flawless. It can also contain a lot of bias” (Code E).
Human journalists and AI algorithms
The RQ2 is about the views of journalists towards the changes in journalism due to the algorithmic process. More than half of the participants had negative views towards automated journalism and they used words like ‘threat’, ‘scary’, ‘insane’ to describe the feelings towards algorithmic process. The insertion of Artificial
Intelligence into the journalistic field creates insecurity about the future of the journalistic field to the majority of the participants. The pessimistic view of
journalists regarding automation is that these algorithms are a potential threat for the field, since they can replace journalists and it can lead to decrease in the journalistic values and make the journalists less respected. “Our field is going through so much hustle already about propaganda and fake news and all that (meaning automated journalism) it is just an extra layer of confusion” (Code B). However, the participant E does not agree that journalists can be replaced, since people in this moment of the time they also believe that the robot journalistic content should not get the opportunity to interview majors or politicians and therefore having humans in control. The
pessimist journalists feel “scared”, “unhappy”, “very sad”, and describe automated journalism as a “not credible” and untrustworthy” machine, which create news. However, the optimistic journalists perceive news-writing programs as an opportunity to make their job easier. The AI algorithms will provide human journalists with a draft that would be able to enrich with their narrative talents. In order to deal with robotic journalism, human journalists should understand that no automated-writing software is able write an opinion column or to become a guardian of democracy.
According to the participant G “I think dangerous is to deny what kind of benefits we could gain with automated journalism”. The inherent limitations of computer
software have encouraged the journalists to pursue cybernetic newsrooms, where journalists and robots will work together and create content, since the robot will be unable to add narrative views and critical thinking into the articles. In order to deal with automated journalism, human journalists should expand their knowledge into the topic. As the journalist E said, “I would be really curious of this process and I would really ask my co-workers who work on automated journalistic projects to get more information on how the technology works”. The negative views towards automation are based on the fact that there is not much information about automated journalism yet, and the human journalists are afraid that a software will take their jobs, but actually it can replace only journalists who do not create context, like journalists who are writing only facts and statistics A journalist who covered the revolution of
Ukraine said, “the first thing I realize when I stepped out of the car was a very
distinctive smell of death, and I can tell you how it smells. It smells like fire, urine and blood” and that’s how this journalist tried to support the argument that a human journalist will never be replaced, because a robot does not have senses and critical thinking. The participants agreed that collaboration is the best solution, in order for humans and robots to co-exist. It is an effective way to help journalists do an in-depth investigation, while an algorithm is offering its services by creating a draft for them.
The role conceptions of journalists
The potential changes that might occur in journalists’ role conception because of automated journalism were the RQ3. The views of journalists regarding their role conceptions were that the roles might change because of the rise of technology and
the competition. The evolution of journalism, urges human journalists to pursue growth and inspire them to learn new techniques of writing. As one of the journalists said “journalists will probably have to start profiling themselves as something different from automated content and social media news...there is no point why um… a journalist should try to compete with a robot, because you already know that a robot can do it better”(Code B). Some of the participants were feeling that their values and the respect towards their role are going to be decreased and their field is already facing so many problems such as fake news, misinformation and bias. They mentioned that some journalists have succeeded in shaping a bad opinion towards the field because of their interests and that has already stigmatized their role conceptions. The manipulation of the news leads the journalists to lose the importance of their role conceptions by becoming less credible and trustworthy in the eyes of their audience. The majority of the participant journalists agreed that the news-writing programs are biased and create even more issues for the role conceptions of the journalists to deal with. Further, they will need to prove that journalism is about people who are using their critical thinking, their narrative skills and putting emotions to their writing with the aim to make the reader understand the significance of the topic. Also, the
importance of being a transparent and a critical thinker, being an investigator and on the ground journalist, who is doing in-depth analysis, will be disqualified because of the automation. The journalists were feeling frustrated and concerned that these values and their role conceptions will disappear, once the algorithms make its integration.
Although software is not sophisticated enough to change the role of a narrative journalist, some of the participants had a different opinion towards the influence of algorithms on their role conceptions. They stated that it will influence their role in a
negative way, since it will appear human journalists as if they are not needed anymore in the newsrooms and this “trend”- as they call it- will take place soon. One of the role conceptions that are under a threat is the role of disseminator. The journalist that has this role is the one that gathers fast information and disseminate them to a large audience. An automated journalist is doing the exact same job, but in a cheaper way and has the ability to generate more stories in a short time than a human journalist can do. One of the participants supported that this role is not as important as the other ones and said “people who are just giving facts in the form of 5Ws and an H and they are not creating any context around that, not doing the work of making people understand why something is important to them, why something matters to their lives personally, how the context of this information will affect them. Frankly, they are not doing a job better than a computer can do. They are not doing much of a job. I think it is a signal maybe you are not doing that interesting job” (Code I). So this job can easily be replaced once the newsrooms decide to install the algorithms. But, the rest role conceptions, such as the interpretive, adversarial and populist mobilizer are not possible to be easily replaced, at least not at the moment. These roles are using their talent to create a context and a story that will appeal to emotions and to logic. The power of having the role of interpretive, popular mobilizer, and adversarial, can be understood at the article written by these kinds of human journalists. These journalists write long reports, full of emotions and facts that have been analyzed in-depth and investigated, with the aim to provide a credible, full of knowledge article to the reader. For example, the journalist who covered the revolution in Ukraine has the ability to analyze and dig deep into governmental issues this is the interpretive role. The journalist who covers such a topic needs to have the ability to describe what is happening in detail and to analyze how people reacted during the evolution. Also, to
transfer to the readers the emotions, that existed in that moment, in order to make him or her realize the seriousness and the tragedy of the revolution. These skills are beyond any of automated journalist, and it is not able to compete and replace such a talent. The limitations of software are raising the hopes of journalists that these kinds of role conceptions will never be “touched” by Artificial intelligence.
Discussion
The purpose of the current study was to determine to what extent does automated journalism affects journalists and their role conceptions. The research has shown that the majority of the participants were not completely aware of what automated journalism is and how it is working, thus the interviews indicated that the opinions of journalists were controversial. Due to the lack of knowledge, some of the interviewees were afraid of the algorithms integration into the media organization and they described it as a potential threat but at the same time they were identifying the advantages that automated journalism adds to the newsrooms and even to their profession. The rest of the interviewees, who worked or knew more about the topic, were completely positive about the progress that journalism is experiencing at the moment. My results are in agreement with Lemelshtrich Latar’s (2015) findings, which showed that, in approximately five to ten year, robots will write 90 percent of the journalistic stories, but these algorithms will benefit journalists by saving them time. Narrative journalism will never be under dangerous position by the technology because these journalists are putting critical thinking, color and depth into their articles. A robotic journalist should not be considered as a potential enemy, since a human journalist has the ability to identify others needs (Lemelshtrich Latar, 2015) and become innovative and creative, while a computer will never succeed to reach
that level. Indeed robotic journalists might be a reason to not hire a new journalist (Lewis et al., 2019) and actually the majority of the journalists were having the same bad feelings for the news-writing programs but they agree that a hybrid collaboration will be the best solution for the coexisting of both in the field, without the one
threatening the other. The interviews show that the role conception of journalists will not change a lot, because there still will be a lot of journalists leading the public discussions. However, some of the role conceptions, like the interpretive role, will be influenced by automated journalism. The identity and the behavior of a journalist are defined by their role conceptions, which shapes their professional experiences. The responses from the participants present that journalists feel the need to protect their roles towards automation and reveal how much they value and respect editorial autonomy (Wu et al., 2019). This study found and agreed with Ombeler, Kuczerany, and Valcke (2015), the news-writing programs are bias, and that concerns journalists about the distortions that might have into their role conceptions and create uncertainty and negative feelings about the future of journalism. A journalist will always be a servant of public opinion leading the public discussions. The professional journalists must obtain innovative ways to promote and disseminate their stories.
Results of this study have to be understood in the context of some limitations since automated journalism is a new topic and it has not be studied much, so it was difficult to find many relevant studies. Due to Covid-19 it was not possible to gather more interviews from journalists who have more experience with automated
journalism and another implication that was faced is the difficulty to find journalists who are aware of the topic. Moreover, the in-depth interview method allows the journalists to mirror their own personal experiences and personal views. A further research could possible focus on the new organizations and journalists who have
experience with automation journalism and they are aware of the benefits or the disadvantages to collaborate with algorithms.
References:
Allen, M. (2017). The sage encyclopedia of communication research methods (Vols. 1-4). Thousand Oaks, CA: SAGE Publications, Inc doi: 10.4135/9781483381411 Carl-Gustav Linden (2017) Decades of Automation in the Newsroom, Digital Journalism,
5:2, 123-140, DOI: 10.1080/21670811.2016.1160791
Carl-Gustav Lindén. (2017). Algorithms for journalism: The future of news work. Journal of Media Innovations, 4(1). https://doi.org/10.5617/jmi.v4i1.2420
Charmaz, Kathy. "Grounded Theory." The SAGE Encyclopedia of Social Science Research
Methods. 2003. SAGE Publications. 24 May. 2009.
Chung, D., & Nah, S. (2013). Media Credibility and Journalistic Role Conceptions: Views on Citizen and Professional Journalists among Citizen Contributors. Journal of Mass Media Ethics, 28(4), 271–288. https://doi.org/10.1080/08900523.2013.826987 Coleman, R., Lee, J., Yaschur, C., Meader, A., & Mcelroy, K. (2018). Why be a journalist?
US students’ motivations and role conceptions in the new age of journalism. Journalism, 19(6), 800–819. https://doi.org/10.1177/1464884916683554
Debatin, Bernhard. 2010. “New Media Ethics.” In Handbuch Medienethik [Handbook Media Ethics], edited by Carsten Brosda and Christian Schicha, 318–327. Wiesbaden: VS Verlag.
Ess, Charles. 2009. Digital Media Ethics. London: Polity Press.
Graefe, A, Haim, M (2016) Human or computer? Whose stories do readers prefer? Available at: http://en.ejo.ch/digital-news/human-computer-stories-readers-prefer
Graefe, A., Haim, M., Haarmann, B., & Brosius, H. (2018). Readers’ perception of computer-generated news: Credibility, expertise, and readability. Journalism, 19(5), 595–610. https://doi.org/10.1177/1464884916641269
Hawksworth, J., Berriman, R., & Cameron, E. (2018). Will robots really steal our jobs? Available at:
https://www.pwc.com/hu/hu/kiadvanyok/assets/pdf/impact_of_automation_on_job s.pdf
Joye, S., Heinrich, A., & Wohlert , R. (2016). 50 years of Galtung and Ruge: Reflections on their model of news values and its relevance for the study of journalism and communication today. CM : Communication and Media , 5–28. doi: doi: 10.5937/comman11-9514
Kent, T. (2015). An ethical checklist for robot journalism . Ethical Journalism Network. Retrieved from https://ethicaljournalismnetwork.org/ethical-checklist-robot-journalism
Konstantin Nicholas Dörr & Katharina Hollnbuchner (2017) Ethical Challenges of Algorithmic Journalism, Digital Journalism, 5:4, 404-419, DOI:
10.1080/21670811.2016.1167612
Lemelshtrich Latar, N. (2015). The Robot Journalist in the Age of Social Physics: The end of Human Journalism? In G. Einav (Ed.), New World of Transitioned Media: Digital Realignment and Industry Transformation (pp. 65-79). New York: Springer International Publishing, Cham. https://doi.org/10.1007/978-3-319-09009-2 Lewis, S., Guzman, A., & Schmidt, T. (2019). Automation, Journalism, and Human-Machine
Communication: Rethinking Roles and Relationships of Humans and Machines in News. Digital Journalism, 7(4), 409–427.
https://doi.org/10.1080/21670811.2019.1577147
M R , A., Sreekumar, A., & Babu, S. D. (2019). Automated Journalism; A Study on Selection of News Stories Based on Individual Priority. International Journal of Innovative Technology and Exploring Engineering (IJITEE), 257–262.
Markus Ojala, & Reeta Pöyhtäri. (2018). Watchdogs, Advocates and Adversaries: Journalists’ Relational Role Conceptions in Asylum Reporting. Media and Communication, 6(2), 168–178. https://doi.org/10.17645/mac.v6i2.1284 Matt Carlson (2015) The Robotic Reporter, Digital Journalism, 3:3, 416-431, DOI:
10.1080/21670811.2014.976412
Monti, Matteo. (2019). Automated Journalism and Freedom of Information: Ethical and Juridical Problems Related to AI in the Press Field.
Neil Thurman, Konstantin Dörr & Jessica Kunert (2017) When Reporters Get Hands-on with Robo-Writing, Digital Journalism, 5:10, 1240-1259, DOI:
10.1080/21670811.2017.1289819
Tal Montal & Zvi Reich (2017) I, Robot. You, Journalist. Who is the Author?,Digital Journalism, 5:7, 829-849, DOI: 10.1080/21670811.2016.1209083
van Dalen, A., de Vreese, C., & Albæk, E. (2012). Different roles, different content? A four-country comparison of the role conceptions and reporting style of political
journalists. Journalism, 13(7), 903–922. https://doi.org/10.1177/1464884911431538
Wölker, A., & Powell, T. E. (2018). Algorithms in the newsroom? News readers’ perceived credibility and selection of automated journalism. Journalism.
APPENDIX
Questionnaire:
The questions that were discussed are the following:
(a) What is the perception regarding the author of automated content? (b) What is the motivation of organizations using automated journalism? (c)Is automated journalism threatening journalists?
(d) How the introduction of automated journalism is changing journalism?
(e) What is the difference of an article written by a journalist and an article written by a robot journalist?
(f) Do you think that the algorithm contains bias or is bias free?
(g) Does this affect the readers in any way? And if yes, what are their reactions? (h) How does the algorithmic process affect the individual journalist?
(i) How does the individual journalist deal with robotic journalism?
