Beyond Debunking: Participatory Fact-Checking in the Age of AI | Caroline Lindekamp

A journalist examines photographs, documents, and source connections through a magnifying glass.

How can societies respond to disinformation when AI makes convincing false content easier to produce? Caroline Lindekamp, Director of CORRECTIV.Faktencheck, argues that automation is not enough. More people must be able to scrutinise claims, understand journalistic standards and participate in verifying information.

In her contribution to the From Prompt to Propaganda workshop in Dresden on 22 September 2026, Lindekamp explained how CORRECTIV.Faktenforum combines professional journalism, citizen participation, structured evidence and research partnerships. This editorial account presents her central arguments without relying on the workshop recording.

An election falsehood as a case study

She opened with an online video suggesting that postal votes for the AfD were being invalidated during the September 2026 state election in Saxony-Anhalt. CORRECTIV investigated the claim and found that the purported ballot papers did not match official requirements. The distribution pattern also suggested a connection to the Russian influence operation Storm-1516. Read CORRECTIV’s documented fact-check.

Beyond detecting fakes

AI changes not only the appearance of misinformation but also its speed, distribution and consumption. As people increasingly use automated assistants to obtain answers, Lindekamp raised a fundamental concern: whether delegating answers also means delegating the critical thinking needed to assess them. The civic-tech approach she presented begins not with what technology can automate, but with what people need in order to participate in a credible information ecosystem.

From a newsroom pipeline to a participatory fact-checking cycle

Professional fact-checking involves monitoring for problematic claims, deciding which are suitable for verification, investigating the evidence, applying editorial standards, publishing a conclusion and reaching audiences through websites, social networks and messaging channels. Lindekamp described how CORRECTIV works across these stages, including crowdsourced tips, manual observation, reverse-image searches, interviews and documentary research. She also highlighted a difficult distribution problem: people who most need a correction may be the least likely to seek out a fact-checking article.

CORRECTIV.Faktenforum opens this workflow to the public. Rather than requiring one participant to prepare a complete fact-check, the platform allows people to contribute manageable pieces of work: reporting a dubious claim, finding its original source, examining a document, discussing an image or helping collect the evidence for a conclusion. The work remains subject to professional editorial scrutiny before publication. This balance between participation and accountability is central to the model.

A community that learns by investigating

Lindekamp said approximately 3,000 people had registered on the Faktenforum platform by the time of the workshop. Participation varies considerably, and the project does not assume that every user will become a full-time volunteer journalist. Its design allows newcomers to start with small investigative tasks and learn from interactions with professional fact-checkers and other contributors.

The wider CORRECTIV fact-checking unit is interdisciplinary: Lindekamp described journalists, community managers, media-literacy trainers and software developers working together. The aim is to offer a genuine editorial process rather than a simplified educational simulation. Participants gain familiarity with journalistic standards by helping apply them, making the CORRECTIV.Faktenforum both an investigative resource and a form of civic education.

She situated this approach within the idea of an editorial society: a society in which many more people can contribute to public debate by understanding and using the methods of responsible journalism. The objective is not to make every citizen a professional reporter, but to increase practical competence in scrutinising sources, evaluating evidence and explaining conclusions.

The platform is also a structured evidence archive

Behind the collaborative interface sits a less visible data project. Investigations connect claims with the places they appeared, the sources collected for and against them, and the evidence that supports an eventual judgment. Lindekamp explained that the platform also serves as part of CORRECTIV’s content-management workflow and builds a structured record of claims and related sources.

This structure makes earlier work reusable. When a similar claim resurfaces in new wording or on another platform, a claim-matching function can help contributors find relevant previous investigations and supporting documents. Lindekamp also pointed to potential benefits for research collaboration and possible adoption or exchange with other European fact-checking organisations. Reuse, however, still requires context-sensitive human scrutiny: a related claim is not necessarily an identical one.

Where AI can help—and where it cannot replace judgment

AI-assisted analysis may help people find related claims, organise material, translate information or navigate investigative steps that would otherwise be difficult for non-specialists. Yet Lindekamp was careful not to present automation as a universal efficiency gain. A tool may fail to save time or fail to meet professional verification standards; either outcome matters when a newsroom must justify every published conclusion.

Her preferred measure of success is broader than throughput. Even when an assistant does not eliminate a research step, it might allow more people to carry out that step competently. The question then becomes whether technology makes evidence and methods more accessible, rather than whether it can produce an answer with no visible human reasoning.

Trust is built in communities, not only online

Lindekamp stressed that the Faktenforum is not solely a software platform. She described bringing established community members together at CORRECTIV’s Berlin newsroom for peer-learning and train-the-trainer sessions, enabling experienced contributors to guide new participants. In-person contact matters because sustained engagement depends on relationships, mutual support and trust.

The CORRECTIV.Faktenfonds extends this logic to local civil society. Community foundations can support events and projects that make fact-checking and media literacy relevant in everyday life. Lindekamp also mentioned collaboration with voluntary fire brigades and exploratory conversations with football clubs. These organisations can reach people who might not engage with fact-checking when approached by a national media brand.

The practical insight is that a correction’s quality and its social reach are different problems. Evidence is essential, but trusted local relationships may determine whether evidence is discussed and taken seriously in the first place.

Beyond a German-language information environment

Another priority is connecting language communities. Lindekamp discussed a move towards multilingual participation, initially involving Russian- and Turkish-speaking communities in Germany. Influence narratives do not travel unchanged between languages: references, audiences, platform use and cultural contexts can differ. Extending fact-checking into these environments therefore requires more than translating completed German-language articles.

The work connects with kuKI, a research collaboration involving CORRECTIV, the FZI Forschungszentrum Informatik, CeMAS and other partners. The project is developing AI-assisted, culturally aware analysis of disinformation across German, Russian and Turkish, together with multilingual support for the Faktenforum. The intention is to make cross-language evidence more accessible and to involve the communities concerned in developing and evaluating the tools. These are active development goals, not claims that the multilingual system has already been perfected.

For researchers, this creates a demanding combination of problems: how to recognise related narratives across different linguistic settings, preserve cultural context, and support inclusive participation while maintaining consistent standards of verification.

Media literacy is not a challenge for young people alone

During the discussion, Lindekamp challenged the tendency to discuss young people’s susceptibility to misinformation without including their perspectives. She pointed to Salon5, CORRECTIV’s youth newsroom, where young people create journalism with the support of professionals. The experience offers a more constructive approach than treating teenagers only as passive recipients of media-literacy lessons.

At the same time, she argued against focusing on younger people to the exclusion of others. Older adults can face different vulnerabilities, including less familiarity with online platforms, and public resilience cannot wait for media education to gradually spread through successive generations. Different audiences require approaches that respond to their particular habits and circumstances.

Responding to accusations of political bias

One audience question concerned accusations that fact-checking organisations apply political standards selectively. Lindekamp acknowledged that such attacks are a substantial challenge, not only for CORRECTIV but for fact-checkers more broadly. Her answer focused on transparency and professional method. A rigorous fact-check makes its sources and reasoning inspectable, identifies what was actually claimed, and gives readers a way to understand how the conclusion was reached.

The Faktenforum adds another form of openness: members of the public can observe and take part in real investigative work instead of merely receiving a finished verdict. Participation does not guarantee that disagreements disappear, but it allows criticism to be directed towards evidence and process. Lindekamp also noted that defending the public value of verification requires support from a wider ecosystem of researchers, educators and institutions, not fact-checking newsrooms alone.

European governance and democratic safeguards

In a final exchange about developments in the United States, Lindekamp urged caution about assuming that Europe should simply follow the US technological trajectory. She acknowledged European dependence on large US technology platforms and the speed of technological change. But she also suggested that rules intended to protect public debate and democratic processes should be evaluated as potential strengths rather than automatically cast as signs of being technologically behind.

The discussion connected the pressures facing fact-checking with wider questions about the governance of digital platforms, the standing of professional journalism and the conditions for open public debate. These questions extend beyond the capabilities of any one AI detection system.

What this approach means for research

Lindekamp’s presentation suggests that evaluating technology for fact-checking requires more than benchmarks for detection accuracy and throughput. Research can also examine whether assistance builds or erodes critical-thinking skills; whether evidence is traceable and reused with its context intact; how citizens and professionals can cooperate without weakening editorial standards; and which multilingual or local designs make participation possible across different communities.

These are analytical implications of the approach discussed in Dresden, not findings that any single platform has already proven. The broader lesson is that technical innovation, journalism and public participation should be designed together. A tool that delivers correct answers but leaves people unable to examine the evidence addresses only part of the challenge.

Resilience to disinformation means more than detecting false content. It also means creating the conditions in which people can understand, investigate and challenge what they encounter.

Further reading

CORRECTIV.Faktenforum · CORRECTIV fact-check: Saxony-Anhalt election video · FZI: the kuKI research project · CORRECTIV.Faktenfonds

This article is an editorial synthesis of Caroline Lindekamp’s contribution to the Dresden workshop on 22 September 2026, supplemented by the linked primary sources. It is not a verbatim transcript.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *