Your questions – our answers

What does SOURCE stand for?

SOURCE stands for „Scalable, Open, and Comprehensive Detection of Disinformation Campaigns on the Web”. The project develops a freely accessible database and infrastructure for detection of disinformation on the web.

When had the project begun and how long does it run for?

SOURCE kicked off on 1st April 2026 and runs over a period of three years.

Who funds SOURCE?

The project is receiving funding of 2,500,000 Euros from the Federal Ministry for Research, Technology and Space (BMFTR) – as part of the funding stream ‘Trust in Democracy and the State: Recognising and Countering Digital Disinformation’ under the ‘Digital. Secure. Sovereign.’ research framework programme.

Which problem does SOURCE solve?

Currently existing approaches to detecting disinformation face two limitations:

To date, the search for disinformation has relied on commercial search engines. However, these are not designed for this type of analysis. They offer only limited capabilities for data filtering and analysing the origin of content – yet both are crucial for the effective detection of disinformation.

Generative artificial intelligence makes it possible to produce misleading and manipulated content faster, more easily, more credibly and on a never before seen scale. AI-generated content is an important indicator of disinformation. Nevertheless, this key indicator of disinformation has so far been largely unusable.

What is the SOURCE-database?

The SOURCE database is a globally unique, freely accessible collection of potential disinformation artefacts from the web and social media. All entries are automatically tagged with disinformation indicators – both in terms of content (e.g. AI detection, argumentative structures) and origin (e.g. dissemination channels, actors involved).

How does SOURCE detect AI-generated content?

The team at the University of Kassel builds on decades of research into digital text forensics. The basic premise is that language models – much like humans – have a recognisable writing style that is difficult to disguise. On this basis, detection methods are developed and evaluated in regular academic competitions (shared tasks).

What is the difference between origin-based and content-based detection?

Origin-based detection (University of Passau) asks: Where does the content come from? Who is disseminating it, when, and via which channels? Content-based detection (University of Kassel) asks: What does it say? Is the text AI-generated, written in a manipulative way, or factually incorrect? Both approaches complement each other and are fed together into the SOURCE database.

Who can use the SOURCE-database?

SOURCE is aimed at a wide range of users: academic communities, journalists and fact-checkers, civil society organisations, as well as businesses and public institutions. Access is open and free of charge.

How can I or my organisation get involved?

SOURCE takes a participatory approach and actively involves specialist communities – for example, in needs assessments, testing of database content and the further development of the tools. Anyone interested can get in touch via the Open Search Foundation or the project partners.

Who is behind SOURCE?

The consortium comprises five partners: the University of Passau (consortium lead, source-based detection), the University of Kassel (content-based detection), the Leibniz Supercomputing Centre (infrastructure), the Open Search Foundation (community building and outreach) and Alliance4Europe (use cases and expert studies). The Text and Data Mining Unit of the European Commission’s Joint Research Centre (JRC) is also involved in a supporting capacity.

Is SOURCE Open Source?

Yes – just as with the OpenWebSearch.eu project, on which SOURCE is technically based, an open-source community is being established around the analysis pipeline. The aim is to collaboratively develop new disinformation indicators and to expand the platform together with the community.

What is disinformation?

Disinformation refers to false or misleading information deliberately disseminated with the aim of deceiving people, manipulating opinions or causing confusion. Unlike misinformation, which is unintentionally false, disinformation is used deliberately – for example, in political campaigns, in the context of crises, or to discredit individuals and institutions.

Why is disinformation a problem?

Disinformation has real-world consequences: people who believe false information about vaccines do not get vaccinated – and diseases spread. Voters who fall for manipulated news make decisions based on false information. Companies or individuals who fall victim to targeted false claims suffer reputational damage that is almost impossible to reverse.

Disinformation is particularly dangerous for democracy: democracy thrives on the free exchange of opinions – but this exchange only works if it is based on a shared set of facts. Disinformation is aimed precisely at destroying this foundation. It sows doubt about institutions, the media and science, fuels mistrust of democratic processes and makes it harder to agree on a shared reality at all. Those who no longer know what is true withdraw – or become susceptible to extreme positions that promise simple answers.

What’s more, a correction usually reaches far fewer people than the original false report – and often comes too late. The rise of AI-generated content is further exacerbating the problem: texts, images and videos can now be faked more quickly and convincingly than ever before, and it is becoming increasingly difficult to distinguish the genuine from the manipulated.

This is precisely where SOURCE comes in: as a tool that helps newsrooms, researchers and anyone else with an interest to identify and assess disinformation at an early stage – before it spreads unchecked.

What happens next? What are the next steps?

Over the coming months, work will begin on the SOURCE database, the crawler system and the first AI-based analysis modules. At the same time, specialist communities and industry partners will be involved in the co-creation process.