How can news media wean themselves off the Big Tech drip?
Researchers and news organisations are building a digital infrastructure that relies less on US tech companies
6 August 2026
Trump’s 2024 return to the White House sharpened Dutch media’s awareness of their dependence on US tech companies, according to Cools. Much of their work is interwoven with companies such as Microsoft, Google, Meta (owner of Instagram and WhatsApp) and OpenAI (maker of ChatGPT). A more independent digital future is vital. Dependence on US Big Tech entails both democratic and economic risks.
Risk to our democracy
Much of organisations’ digital infrastructure—servers, cloud systems and software—is controlled by a small number of US companies. External commercial parties such as Google, Microsoft, Meta and OpenAI can therefore influence access to information. This leaves organisations with little control of their own.
One of the judges there suddenly lost access to his email account because Microsoft had blocked him.’
A judge at the International Criminal Court in The Hague suddenly lost access to his email account after Microsoft blocked him
‘Last year you saw the consequences of this at the International Court of Justice in The Hague. One of the judges there suddenly lost access to his email account because Microsoft had blocked him.’
This example comes from the judiciary—another pillar of our democracy—but it could also occur in journalism. Developments like these increasingly undermine the watchdog role journalism plays in a democracy.
Risk to news organisations
You’ve probably noticed it yourself: search engines and platforms are building all kinds of ‘AI layers’ on top of online search results. That may be convenient for users, but it also means fewer people click through to news organisations’ sites. Social media contribute to ever‑lower referrals to news sites too, simply because platforms like Meta focus on keeping users within their own ecosystem. The expectation is that visitor numbers to news sites will drop sharply in the coming years.
Hannes Cools is Assistant Professor of Communication Science at the UvA and affiliated with the AI, Media & Democracy Lab. In his research he combines theory and practice: he has previously worked with newsrooms such as the BBC and the NOS and with major media companies like DPG and Mediahuis, examining how journalists use AI in their daily work and what concerns they have. Cools has also worked as a journalist himself.
Towards greater digital autonomy
According to Cools, newsrooms are beginning to realise they currently lack sufficient digital autonomy. That means, for instance, that they do not control their own data storage because it is in the hands of US companies. The project Cools is now involved in is an important step towards greater digital autonomy. The aim is to move away from US Big Tech without bringing these functions in‑house: instead, the team will identify European and open‑source providers that better align with the news organisations’ values.
Collaboration with research partners and the field
‘From the AI, Media and Democracy Lab at the UvA, we are working with a range of partners on this project. For example, with the Faculty of Law, because of their legal expertise in regulation, infrastructure and licences. The Amsterdam University of Applied Sciences (AUAS) also plays a role, as they have strong technical expertise in open‑source language models.’ But for Cools the practice partners are at least as important: ‘We are working with the think tank Waag, and with De Correspondent and Follow the Money (FTM). These media organisations have a strong need for digital autonomy. The grant for this project goes not only to the UvA but also to them, which creates even more motivation to work closely together.’
Research design: from definition to prototype
The project will run for around two years from 2026 and will consist of three components:
- The researchers will explore with the practice partners what they themselves mean by digital autonomy. They will also map the ways they currently depend on Big Tech and what alternatives are available; for example European platforms or open‑source language models.
- The team will then engage the public through focus groups. Participants will vary in age, experience with AI, and whether or not they are regular readers of FTM or De Correspondent. The focus groups will discuss their definition of digital autonomy and their expectations of what AI should be able to do in a news context.
- In the final phase of the project, the news organisations will build prototypes of features for readers to test. Readers will use the newly selected or purpose‑built AI models to, for example, search the organisations’ archives, have articles summarised, or listen to stories in podcast form.
Over time, we hope that larger news organisations such as DPG will also take steps towards greater digital autonomy
What will readers notice?
As Cools envisages it, readers will, in time, mainly notice that the use of AI in journalism becomes fairer and more transparent. The hope is that it will also become easier to search the archive, have articles summarised or read aloud. ‘And in the longer term we hope, of course, that larger news organisations like DPG will also take steps towards greater digital autonomy, so that journalism as a whole truly retains control over its own work and its role as the democracy’s watchdog.’