Читать книгу Rethinking Digital Native Comunicators Training - Cristina Pulido Rodriguez - Страница 8

Is Artificial Intelligence becoming my Coworker? How the Role and Tasks of Journalists Change Through the Use of Artificial Intelligence

MARCEL FRANZE Ostfalia University of Applied Sciences

Keywords: Artificial Intelligence, AI, Journalism, Newsroom, Role

Abstract: The arrival of artificial intelligence (AI) in journalism is creating pressure for change. AI is able to take over tasks and roles within the newsroom that were previously filled by human editors or performs tasks that could not be done by humans before. This paper presents the results of a survey of 13 journalists and media experts from Germany. The survey aimed to find answers to the overarching question: How do the role and the task of journalists and their function in the editorial office change through the use of artificial intelligence? It turned out that there is a clear expectation towards AI in the editorial office. AI programmes should be tools that take away tedious work, do not lead to job loss and reduce workload. The perspective of AI becoming a colleague is none shared by those working in this field.

According to the Media Worker Report, seven out of ten “media workers” have already been influenced by digitalisation in crucial processes of their work in the last five years. One-third of respondents expect AI to have an impact on their job. 47% expect “somewhat” and 23% expect “very much” an influence of technologies such as Big Data and AI on their job profile (cf. nextMedia 2018). So far, AI systems mainly generate shorter texts in the areas of sports reporting, weather, finance, stock exchange and traffic. These AI systems are based on text modules that need large databases and good data quality. For example, AIs have been writing reports for the news agency AP, the Los Angeles Times and Washington Post for several years. Although, longer texts are also possible. Even a whole scientific book written by an AI is published already (cf. Writer, 2019). In terms of the quality of the texts, studies have found that differences between human and automated written texts are almost indiscernible (cf. van der Kaa, & Krahmer, 2014; cf. Clerwall, 2014). However, as Reinhard Karger, the company spokesperson for the German Research Centre for Artificial Intelligence (DFKI) notes, this also leads to uncertainty among journalists, especially about concerning job security (cf. Fischaleck, 2018). The consequence of these developments is that artificial intelligence can and will take over tasks and roles within the newsroom that were previously filled by human editors or perform tasks that cannot be done by humans.

RESEARCH QUESTION AND RESEARCH DESIGN

This section presents the results of a qualitative survey of journalists and media experts. The overarching research question was: How do the role and the task of journalists and their function in the editorial office change through the use of artificial intelligence? Four key points were of interest: First, for what purpose is AI being used in the newsroom? Secondly, what role does AI take on in distinction to the role of journalists in the newsroom? Thirdly, what specific tasks does AI take on in the newsroom and what tasks remain with journalists? Fourthly, what are the future prospects and scenarios regarding the tasks and roles of AI in the newsroom? The empirical basis is the data collection using guided interviews and the subsequent qualitative content analysis. A total of 13 interviews were conducted with experts from Germany. The telephone interviews were between 30-45 minutes long and took place between November 2019 and January 2020. The interviewees were chosen deliberately. The decisive factor was their role as an expert. A person was considered an expert if he or she worked for a journalistic media company and was either a journalist or a contact person for new technologies in the media company. Further, a person was considered an expert if they had knowledge of journalistic work routines or were able to assess major developments in journalism. Typical job titles of the participants were project staff for AI, editor-in-chief and editor. Nine of the 13 interviewees were employees of media companies publishing a regional media outlet and were located all over Germany. The remaining four were either self-employed or worked nationwide.

It was already anticipated in advance of the survey that it would be challenging to find editorial offices that use AI. For this reason, the survey was deliberately opened up to editorial offices that do not or do not yet use AI. A differentiation of AI, for example into categories such as weak and strong AI (cf. Frankish; Ramsey 2014: 342) or symbolic and sub-symbolic AI (cf. Sieber 2019: 61) was deliberately not made. The term AI was defined broadly: Artificial intelligence is a collective term for various technologies and software applications. Based on algorithms, it attempts to virtually emulate the performance of the human mind or even to expand and optimise it. It is supposed to be able to autonomously solve complex problems. Alternative questions were asked for respondents who do not use AI in their newsroom. For example, if a company or an editorial department does not yet use AI, the question of when AI was used for the first time is not necessary. However, it was at least possible to ask what the reasons for non-use were and whether the use of AI is planned for the future. Thus, the questionnaire was designed as a decision tree scheme. The guided interviews were evaluated by means of a qualitative content analysis. The evaluation was divided into the following thematic complexes: Purpose of the use of artificial intelligence, role of AI vs. role of journalists, tasks of AI vs. tasks of journalists as well as future forecasts and scenarios.

RESULTS

As previously expected, only a minority use AI programmes. Only four of the 13 interviewees use AI in their newsrooms. Six do not use AI yet and do not plan to do so in the longer term. Two have planned the use of AI in a very timely manner. The purposes of using AI in the newsroom are: Screening social media (filter), automated text generation (e.g. reporting about elections), identification of hate speech on social media. The respondents name efficiency and curiosity about new technical developments as overriding reasons for use. The reasons given for not using AI solutions are that they are simply unnecessary or that the staff is lacking the necessary know-how.

The role that AI takes on in the newsroom or how the respondents imagine the role of AI in the newsroom (for respondents who do not use AI) is described as: “supporting role” (int. 2), “service” (int.8) “not a full-fledged editor” (int.1) or as “unskilled labour, assistant” (int. 10) which “does the unpopular tasks” (int. 4). The tasks that AI takes on in the newsroom or how the respondents imagine the tasks of AI in the newsroom (for respondents who do not use AI) are first and foremost support activities. Examples are automated data-supported text generation as well as the creation of interview transcriptions or short press releases for the domains of traffic, sports, weather, stock exchange, and real estate. These are all areas that can be described as ‘highly standardised reporting’. AI is said to have a particular strength in “determining user needs and the individualised playout of content” (int. 1). According to the respondents, a central task that remains and should remain with the journalist is the stimulation and formation of public opinion. Furthermore to remain are the selection and compilation of topics and the writing of stories, as illustrated by a quote: “Cannot imagine AI writing comments, as it lacks the feelings and life experiences to do so” (int. 4). Somewhat surprising is the mention of research, which some respondents believe should remain in the hands of journalists. This is actually a task for which AI is well-suited, for example in sifting through large amounts of data.

Asking about how the role and tasks of AI in the newsroom will develop in the future, the concern about the workplace takes up a large part. Two camps formed. Some of the respondents do not see AI threatening jobs. Younger, tech-savvy people, in particular, see AI as an opportunity, less as a threat. The other camp, however, is very much worried about losing their jobs or has the impression that older colleagues are worried about losing their jobs. One respondent can imagine that jobs at the news desk will disappear. Detached from the topic of job security, the respondents believe that AI cannot replace journalists in the long term but can only take over individual aspects of their role and work. A positive expectation of the respondents is that AI could free up journalists’ capacities by taking over unpopular and standardised work, thus reducing the workload.

The findings of the survey indicate that AI programmes have not yet had a drastic impact on the role and task of journalists in the newsroom. This is mainly due to the fact that only a minority of the interviewed journalists and media experts use AI programmes. The newsrooms that do use AI programmes, do so on a limited scale and use them for only one purpose or task. Nevertheless, the expectation towards AI in the newsroom is relatively clear. Respondents do not see or want the role of AI to go beyond assistance. Ideally, AI programmes should be tools that take away tedious work, do not lead to job loss and reduce workload. The perspective of AI becoming a colleague is none shared by those working in this field.

REFERENCES

Writer, B. (2019). Lithium-Ion Batteries. Springer International Publishing. https://doi.org/10.1007/978-3-030-16800-1.

Clerwall, C. (2014). Enter the Robot Journalist. Journalism Practice, 8(5), 519–531.

Fischaleck, F. (2018). Künstliche Intelligenz: „Eine unglaubliche Zeitersparnis für Journalisten“. https://www.fachjournalist.de/kuenstlicheintelligenz-eine-unglaubliche-zeitersparnis-fuer-journalisten/.

Frankish, K., & Ramsey, W. M. (Eds.). (2014). The Cambridge handbook of artificial intelligence. Cambridge University Press.

nextMedia.Hamburg. (2019). So denken die Deutschen über Künstliche Intelligenzen: Vertrauen in KIs steigt deutlich an. https://www.nextmedia-hamburg.de/so-denken-die-deutschen-ueber-kuenstliche-intelligenzen/.

nextMedia.Hamburg (2018): #dasmitmedien. Media Worker Report. https://www.nextmedia-hamburg.de/wp-content/uploads/2018/12/dasmitmedien-Media-Worker-Report-2018.pdf.

Sieber, A. (2019). Dialogroboter: Wie Bots und künstliche Intelligenz Medien und Massenkommunikation verändern. Springer Fachmedien Wiesbaden GmbH.

van der Kaa, H. A. J. & Krahmer, E.J. (2014). Journalist versus news consumer: The perceived credibility of machine written news. https://research.tilburguniversity.edu/files/4314960/cj2014_session4_paper2.pdf.