The way news reaches us is changing. AI systems are increasingly becoming gateways to information, allowing people to find news without visiting the original source. This gives large tech companies greater influence over how news is distributed, putting the independence of news media and the diversity of the media landscape under pressure. Waag Futurelab investigates how European and open-source alternatives to big-tech language models (LLM's) can contribute to greater digital autonomy.
When is an alternative good enough?
European and open-source technology can make news media less dependent on large tech companies. But where a technology comes from does not necessarily determine whether it works for a news organisation. It also needs to function well, fit journalistic practice and reflect values such as reliability and independence.
Waag therefore investigates what digital autonomy means for news media in practice. When does technology actually contribute to greater autonomy? What new dependencies might emerge? And how much freedom does a news organisation retain to decide how technology is used?
Within the Digital Autonomy that Fits project, Waag investigates these questions together with the University of Amsterdam, Amsterdam University of Applied Sciences, De Correspondent and Follow the Money. The project focuses on European and open-source large language models and how this technology could support news media in distributing their content and reaching their audiences. Together with newsrooms and news readers, Waag investigates how these language models should be designed and which values should underpin their development.
European language models in practice
The project partners will test existing open-source language models at De Correspondent and Follow the Money. As part of this work, they are developing an interface that allows readers to search the digital archives of both news organisations. This helps reveal what is needed to use European and open-source language models within a news organisation in practice.
This involves more than technical performance. Waag also investigates which choices news organisations can continue to make for themselves, which dependencies the technology creates and what different technological choices mean for newsrooms and readers. It is this combination that determines whether an alternative not only works technically, but also fits the practices and values of a news organisation.
Newsrooms and readers help shape the technology
Waag involves newsrooms and news readers in the research from the outset. In co-creation sessions, Waag researchers work with them to identify the needs and values that matter when using language models for news and translate these insights into design principles for the technology. In this way, newsrooms and readers play an active role in shaping what a suitable alternative should look like.
Waag incorporates these principles into the testing protocol used to evaluate the selected language models with newsrooms and readers. This means looking not only at how the technology performs, but also at the values and needs identified earlier in the research. The research therefore helps reveal what is needed to use language models in ways that fit journalistic practice.
Involving newsrooms and readers is part of Waag's broader research into public technology. Waag explores how digital technology can be developed from public values and how people can have a say in its design and use. From this perspective, newsrooms and readers in Digital Autonomy that Fits do not simply assess a technology once it has been developed. They are involved from the outset in shaping what a suitable alternative should be able to do.
Digital autonomy for the wider media sector
The insights from the research should help news media make more informed choices about the technology they use and the dependencies that come with it. Waag and the project partners are therefore developing a framework that news organisations can use to assess different technological solutions from the perspective of digital autonomy. In this way, the lessons from the research can also be applied across the wider media sector.
Digital autonomy is not simply about switching from one technology to another. It is about enabling news organisations to decide for themselves which technologies they use, under what conditions and which parties they depend on in the process. By designing and testing alternatives together with newsrooms and readers, Waag investigates what news media need to make those choices based on their own practices and values.