Flexit Lindholmen Science Park: The automation of journalism. Innovation, collaboration and knowledge management when implementing AI in news organizations
The aim of this project is to understand how news organizations work to manage various competencies, work flows and (sometimes) conflicting goals in the process of automating journalism.
Based in the intersection of two uniquely collaborative platforms, Medier & demokrati and AI innovations of Sweden, the project seeks to understand three entwined aspects: 1) the network of actors involved in AI innovation processes, 2) the meaning and methods for negotiation and collaboration between actors of those networks and 3) the meaning of organizational settings for building collaborative and innovative newsrooms. The point of focusing those key themes is to create understanding of how fundamental journalistic values may be shaped and enacted in the automation processes of news work.
Empirical case studies will be drawn from a selection of Swedish new organizations where AI is currently being implemented. The study applies an inductive ethnographic approach including interviews with actors involved in technical innovations (journalists, developers, sales people, editors etc.), observations of collaborative meetings and work processes, and document research of strategic and policy documents, annual reports etc.
The theoretical framework relies on organizational and management theory as well as journalism studies, forwarding the concepts of automated journalism, knowledge management and collaboration.
Final report
Project purpose and development
The aim of the project was to understand how news organizations work to manage different competencies, workflows and (sometimes) conflicting goals in the process of automating journalism. When the project began, automation was still in an initial phase in the newsrooms. The pandemic also meant that the study had to rely on methods other than the ethnographic participation that was intended, and instead interviews and a survey were used, which still worked satisfactorily. Since the AI ??development was above all visible in the parts of the news companies that deal with market, analysis and distribution, the study came naturally to focus on the organizational perspective and leadership in relation to innovation. How do news organisations work strategically with developing and integrating AI? Which actors are involved and how do the processes work? It is these and similar questions that have been addressed in the project.
The implementation
Within the framework of the project, around forty qualitative interviews have been conducted, as well as a large number of conversations and meetings to understand the context. The interview material includes editors-in-chief, strategists, developers, analysts, product owners and others who work with developing AI-driven solutions connected to the news business. Ten of the interviews were conducted with startups and entrepreneurs who deliver AI solutions to the news industry. The material also includes participation in digital meetings linked to the development of a new AI-driven product on the Altinget magazine. These observations provided an understanding of the process and the different competencies that collaborate in such development projects. In addition to the qualitative material, two surveys were also conducted, one in 2021 and one in 2024. The survey was sent to selected media managers in the Nordics and was about taking the pace of AI development in Nordic media companies and mapping issues relevant to the purpose of the study.
How has the project contributed to increased cooperation between the university and organizations outside the academy?
As it is a Flexit project, collaboration has been a central part of the work. It has been an important goal to inject scientific perspectives and networks into Medier & Demokrati's activities, and conversely to strengthen the societal role of JMG and the University of Gothenburg, especially in relation to the media industry. Examples of collaboration spaces that the Flexit era has resulted in:
• Project management of the start-up year of the Future of Democracy initiative, in collaboration with AI Sweden. https://www.futureofdemocracy.se/
• Collaboration between JMG and Media & Democracy in the Erasmus+ funded project Crossborder Journalism Campus, CJC. The project is about developing innovative teaching methods in cross-border investigative journalism. https://www.gu.se/en/study-gothenburg/crossborder-journalism-campus
• Application in 2022 to the Creative Europe Program (CREA) under the title Investigative Journalism AI. The envisaged innovation project involved four universities, three innovation hubs and five newsrooms in a total of eight European countries and aimed to strengthen investigative journalism with AI. Unfortunately, the application was not granted, but the network remains and has led to new collaborations later.
• Application in 2023 to Vinnova by Democreate, a project aimed at the Impact innovation call with Lindholmen Science Park and the University of Gothenburg as principals. The application involved a cross-disciplinary consortium of over 60 actors and aimed to strengthen the creative industry in Sweden. https://www.lindholmen.se/sv/democreate
• Based on my research, I participated in the program group that organized the pre-conference to the international conference Global Investigative Journalism Conference in Gothenburg in September 2023. https://gijc2023.org/pre-conference-day/
• JMG is currently developing leadership training for digital newsrooms in which key people in the Flexit network participate. The training will be given during the academic year 23/24 and is largely based on insights from the Flexit project.
• The project has also led to the development of learning materials and an emerging research network around AI and journalism training.
The project's most important results
1. The leadership has a decisive importance for the innovation climate in the organization and for the extent to which one dares to invest in the development of new data-driven solutions. But the leadership in the news companies often lacks knowledge and interest in AI. This means that the greatest responsibility for the development of AI is placed on technical and product-oriented personnel, and that the journalistic perspective is not always represented as it should be. It also means, conversely, that innovations developed by technical staff may have difficulty gaining impact because the work is not sufficiently integrated and anchored in the editorial work as a whole.
However, the winds have turned with the launch of the major language models in 2023. Nowadays, most companies see the value of AI and zealots who previously seemed quite autonomous have now been elevated to the management teams.
2. Innovation aimed at AI and automation is conducted just as much, if not more, outside the media houses. A conclusion from the project is that today's journalism is produced in an ecosystem where the journalistic profession works side by side with technical, analytical and product development competencies. There is thus reason to reconsider the journalist-centered view that traditionally exists within journalism research. The news is now produced in collaboration and on equal terms. In some cases, it can even be said that journalism has lost ground. Technical and commercial logics not only challenge the journalistic/professional but are in many situations dominant.
3. A large part of the innovation in the media industry is driven by startups, entrepreneurs and larger companies that deliver tech solutions to the news companies. A partial study deals with these actors, who they are and what they do. Their motives, conditions and power of innovation are of great importance for the development of journalism and therefore form an important part of the ecosystem. These interviews confirm that journalism arises in the space between a variety of actors, many of whom have no journalistic background and where the media industry is just one of many industries that tech companies work against. They also show that both tech providers and news companies conduct operations conditioned by large global platforms and cloud solutions. This digital infrastructure seems homogenizing across the media sector, especially as AI builds on scalability and big data.
New research questions
The project deals with an extremely mobile area, so there is a constant need for new knowledge. Examples of questions that can be asked based on the project's results:
• Where in the organization should the innovation work be placed so that it can best be integrated into the business in a sustainable way?
• How has the introduction of major language models affected the dynamics between journalistic and technical/product-oriented teams within news organizations, both in terms of collaboration and balance of power?
• What strategies and collaboration models are news organizations using to leverage external technology providers and startups in AI and automation, while maintaining journalistic values ??and integrity?
• How does the presence of AI competence in management teams affect changes in the innovation climate and the strategies for developing data-driven solutions within news companies?
• How do journalism education programs see their role in relation to the widespread automation that their students will face in the future, both in journalism and in society at large?
Dissemination of research results
Results from the Flexit project have been presented in a number of different contexts:
• Research group DataPublics, University of Copenhagen 230201
• The authority for press, radio and television 230512
• Research network GATE, Chalmers 230516
• Media City Bergen Future Week 220608
• AI-inspiration @ SR, Swedish Radio online 220217
• Media and democracy webinar 220210
• Bodenforsseminariet, School of Economics Gothenburg 220511
• Bokmässan, Gothenburg 220924
• The MIK network Gothenburg, Lindholmen Science park 211124
• EUTOPIA week online 210414
• JMG researcher day 200204
• GAMI Wan-Ifra Media Lab Days, Copenhagen 20200218
The aim of the project was to understand how news organizations work to manage different competencies, workflows and (sometimes) conflicting goals in the process of automating journalism. When the project began, automation was still in an initial phase in the newsrooms. The pandemic also meant that the study had to rely on methods other than the ethnographic participation that was intended, and instead interviews and a survey were used, which still worked satisfactorily. Since the AI ??development was above all visible in the parts of the news companies that deal with market, analysis and distribution, the study came naturally to focus on the organizational perspective and leadership in relation to innovation. How do news organisations work strategically with developing and integrating AI? Which actors are involved and how do the processes work? It is these and similar questions that have been addressed in the project.
The implementation
Within the framework of the project, around forty qualitative interviews have been conducted, as well as a large number of conversations and meetings to understand the context. The interview material includes editors-in-chief, strategists, developers, analysts, product owners and others who work with developing AI-driven solutions connected to the news business. Ten of the interviews were conducted with startups and entrepreneurs who deliver AI solutions to the news industry. The material also includes participation in digital meetings linked to the development of a new AI-driven product on the Altinget magazine. These observations provided an understanding of the process and the different competencies that collaborate in such development projects. In addition to the qualitative material, two surveys were also conducted, one in 2021 and one in 2024. The survey was sent to selected media managers in the Nordics and was about taking the pace of AI development in Nordic media companies and mapping issues relevant to the purpose of the study.
How has the project contributed to increased cooperation between the university and organizations outside the academy?
As it is a Flexit project, collaboration has been a central part of the work. It has been an important goal to inject scientific perspectives and networks into Medier & Demokrati's activities, and conversely to strengthen the societal role of JMG and the University of Gothenburg, especially in relation to the media industry. Examples of collaboration spaces that the Flexit era has resulted in:
• Project management of the start-up year of the Future of Democracy initiative, in collaboration with AI Sweden. https://www.futureofdemocracy.se/
• Collaboration between JMG and Media & Democracy in the Erasmus+ funded project Crossborder Journalism Campus, CJC. The project is about developing innovative teaching methods in cross-border investigative journalism. https://www.gu.se/en/study-gothenburg/crossborder-journalism-campus
• Application in 2022 to the Creative Europe Program (CREA) under the title Investigative Journalism AI. The envisaged innovation project involved four universities, three innovation hubs and five newsrooms in a total of eight European countries and aimed to strengthen investigative journalism with AI. Unfortunately, the application was not granted, but the network remains and has led to new collaborations later.
• Application in 2023 to Vinnova by Democreate, a project aimed at the Impact innovation call with Lindholmen Science Park and the University of Gothenburg as principals. The application involved a cross-disciplinary consortium of over 60 actors and aimed to strengthen the creative industry in Sweden. https://www.lindholmen.se/sv/democreate
• Based on my research, I participated in the program group that organized the pre-conference to the international conference Global Investigative Journalism Conference in Gothenburg in September 2023. https://gijc2023.org/pre-conference-day/
• JMG is currently developing leadership training for digital newsrooms in which key people in the Flexit network participate. The training will be given during the academic year 23/24 and is largely based on insights from the Flexit project.
• The project has also led to the development of learning materials and an emerging research network around AI and journalism training.
The project's most important results
1. The leadership has a decisive importance for the innovation climate in the organization and for the extent to which one dares to invest in the development of new data-driven solutions. But the leadership in the news companies often lacks knowledge and interest in AI. This means that the greatest responsibility for the development of AI is placed on technical and product-oriented personnel, and that the journalistic perspective is not always represented as it should be. It also means, conversely, that innovations developed by technical staff may have difficulty gaining impact because the work is not sufficiently integrated and anchored in the editorial work as a whole.
However, the winds have turned with the launch of the major language models in 2023. Nowadays, most companies see the value of AI and zealots who previously seemed quite autonomous have now been elevated to the management teams.
2. Innovation aimed at AI and automation is conducted just as much, if not more, outside the media houses. A conclusion from the project is that today's journalism is produced in an ecosystem where the journalistic profession works side by side with technical, analytical and product development competencies. There is thus reason to reconsider the journalist-centered view that traditionally exists within journalism research. The news is now produced in collaboration and on equal terms. In some cases, it can even be said that journalism has lost ground. Technical and commercial logics not only challenge the journalistic/professional but are in many situations dominant.
3. A large part of the innovation in the media industry is driven by startups, entrepreneurs and larger companies that deliver tech solutions to the news companies. A partial study deals with these actors, who they are and what they do. Their motives, conditions and power of innovation are of great importance for the development of journalism and therefore form an important part of the ecosystem. These interviews confirm that journalism arises in the space between a variety of actors, many of whom have no journalistic background and where the media industry is just one of many industries that tech companies work against. They also show that both tech providers and news companies conduct operations conditioned by large global platforms and cloud solutions. This digital infrastructure seems homogenizing across the media sector, especially as AI builds on scalability and big data.
New research questions
The project deals with an extremely mobile area, so there is a constant need for new knowledge. Examples of questions that can be asked based on the project's results:
• Where in the organization should the innovation work be placed so that it can best be integrated into the business in a sustainable way?
• How has the introduction of major language models affected the dynamics between journalistic and technical/product-oriented teams within news organizations, both in terms of collaboration and balance of power?
• What strategies and collaboration models are news organizations using to leverage external technology providers and startups in AI and automation, while maintaining journalistic values ??and integrity?
• How does the presence of AI competence in management teams affect changes in the innovation climate and the strategies for developing data-driven solutions within news companies?
• How do journalism education programs see their role in relation to the widespread automation that their students will face in the future, both in journalism and in society at large?
Dissemination of research results
Results from the Flexit project have been presented in a number of different contexts:
• Research group DataPublics, University of Copenhagen 230201
• The authority for press, radio and television 230512
• Research network GATE, Chalmers 230516
• Media City Bergen Future Week 220608
• AI-inspiration @ SR, Swedish Radio online 220217
• Media and democracy webinar 220210
• Bodenforsseminariet, School of Economics Gothenburg 220511
• Bokmässan, Gothenburg 220924
• The MIK network Gothenburg, Lindholmen Science park 211124
• EUTOPIA week online 210414
• JMG researcher day 200204
• GAMI Wan-Ifra Media Lab Days, Copenhagen 20200218