Digital Autonomy that fits: AI autonomy for news media
Project
Media outlets are increasingly using AI tools and cloud infrastructures developed by a small number of, mainly American, technology companies. This dependence affects one of journalism’s core functions: the independent distribution of trustworthy information to the public. The Digital Autonomy that fits project investigates how media organisations can use open-source technology and European language models to gain better control over their content, distribution channels and relationship with their audience. This project is funded by SIDN.
Cause
Generative AI is transforming the way journalistic stories are produced, distributed and consumed. When it comes to innovation, news organisations often rely on the same technology companies with which they are simultaneously competing for the public’s attention. Furthermore, these companies use journalistic content to train AI systems that are increasingly acting as alternative sources of information.
This creates new dependencies that affect not only the business model of media outlets, but also their social and democratic role. As decisions regarding access to information become increasingly centralised among large technology platforms, the diversity of the media landscape comes under pressure and journalism’s independent watchdog function becomes more vulnerable. At the same time, there is a growing realisation in the Netherlands and the rest of Europe that a transition towards greater digital autonomy is necessary. This project investigates how media organisations can take this step.
Goal
The aim of the project is to investigate how news organisations can use open-source technologies and European Large Language Models (LLMs) to gain greater control over the distribution of their content.
In addition, the project is investigating exactly what digital autonomy means for the media sector. What technological dependencies are at play? What alternatives are available? And under what organisational, legal, technical and societal conditions can media organisations achieve greater autonomy?
Approach
Within the project, researchers from the AI, Media & Democracy Lab at the University of Amsterdam and Amsterdam University of Applied Sciences are collaborating with Waag Futurelab, De Correspondent and Follow the Money. Together, they are developing and testing existing open-source and European AI solutions, including GPT-NL. These technologies are being used to develop new interfaces that give citizens better access to the digital archives of De Correspondent and Follow the Money. The exact form is being explored in collaboration with editorial teams and users and could, for example, take the form of a chatbot, widget or other search and interaction tool.
The project uses an action research approach in which research, development and practical application are closely intertwined. Citizens and professionals from the media sector are also involved through participatory design workshops and focus groups.
In addition to the technical development, the project is investigating which social, organisational, institutional, legal and ethical factors support or, conversely, obstruct the transition to greater digital autonomy.
Aimed results
The project will provide practical knowledge and tools to support media organisations in their transition towards greater digital autonomy. Open-source and European AI solutions will be tested, implemented and, where necessary, further refined for journalistic applications.
In addition, the project is developing a vision for digital autonomy within the media sector and a framework for defining, evaluating and putting autonomy into practice. This will result, among other things, in a roadmap that provides insight into dependencies, choices and trade-offs regarding technology, infrastructure and governance.
The knowledge gained is translated into practical guidelines for the selection, implementation and procurement of AI solutions within the media sector. Recommendations are also being developed for policymakers and the Dutch ICT sector.
By giving media organisations greater control over the distribution of their content and by making alternatives to Big Tech visible and usable, the project contributes to a pluralistic media landscape, public values and democratic resilience.
Research partners
- University of Amsterdam (AI, Media & Democracy Lab) – project lead
- Amsterdam University of Applied Sciences (AI, Media & Democracy Lab)
- Waag Futurelab
- De Correspondent
- Follow the Money
Funding
SIDN