Imagine scrolling through your feed and watching a video of a political leader saying something they never said. You open an article from an outlet you don’t recognize, and it feels almost real, except for a few details that don’t add up. This is the new reality of digital life, where synthetic content, manipulated media, and algorithmically amplified falsehoods are no longer crude and easy to spot. They are sophisticated, personalized, and designed to exploit the very technologies we use to stay informed. The threat of foreign information manipulation and interference, or FIMI, has become one of the defining challenges of our time, and democratic societies are scrambling to respond. That is why the European Union has invested heavily in tools to detect and counter disinformation. But a crucial problem remains: these tools have been developed in isolation, tested in controlled conditions, and rarely compared against one another in real-world settings. In fact, there has never been a systematic European effort to ask the most basic but important question: which countermeasures actually work when it really matters? This is the gap that a major new research project, FIMI-RESIST, has been designed to fill.
Led by Professor Vera Schmitt at Johannes Gutenberg University Mainz, FIMI-RESIST brings together a consortium of ten partners from nine countries, uniting researchers, technologists, media professionals, and civil society organizations. The project’s official name is “Foreign Information Manipulation and Interference – Robust European Safeguards and Innovative Systems against Threats,” and it is being supported by Horizon Europe, the European Union’s flagship research and innovation program, with about EUR 3.5 million in funding. Over the next 30 months, the consortium will not build yet another shiny new tool, even though that would be the easier and more familiar path. Instead, it will do something more difficult and arguably more valuable: evaluate, compare, and harmonize the counter-disinformation tools that already exist. Twenty previously funded European tools will be put through a common battery of tests, using shared benchmarks and success criteria. The goal is not to crown a single winner, but to generate robust, honest evidence about which tools are best suited for which users, which operational contexts, and which evolving threat scenarios. By treating the existing landscape of anti-disinformation technology as one ecosystem rather than a collection of stand-alone projects, FIMI-RESIST aims to help Europe move from scattered efforts to a coherent, credible defense.
The project is grounded in the belief that the best way to understand whether a tool works is to watch it in action. That means leaving the sterile conditions of the laboratory and going to the messy, high-pressure environments where journalists chase deadlines, fact checkers triage a flood of dubious claims, and security analysts must decide quickly what is a threat and what is just noise. FIMI-RESIST will work directly with journalists, fact checkers, security stakeholders, and civil society organizations to see how these tools integrate into everyday workflows, where they slow people down, where they create false confidence, and where they simply miss the new tricks made possible by generative AI. This is not only a technical exercise; it is also a human one, because the best algorithm in the world is useless if a stressed editor cannot understand its output or if an independent fact checker cannot trust its reasoning. The project will also establish a help desk, a European hub, and training offerings so that validated methods and lessons learned do not disappear when the funding ends. The objective is to leave behind not just reports and academic papers, but a practical, living infrastructure that can continue to support the people on the front lines of the information war.
None of this happens in a political vacuum. The way Europe responds to disinformation is already being shaped by a wave of major legal and regulatory frameworks, including the Digital Services Act, the AI Act, and the European Media Freedom Act. These laws create obligations and expectations around platform accountability, algorithmic transparency, and media diversity, but they cannot be effective without reliable, independent, and practical tools to support enforcement and compliance. FIMI-RESIST intends to feed its findings directly into this policy debate. By providing evidence about what works, the project can help regulators, lawmakers, and institutional actors make smarter decisions about what to require, what to fund, and what to expect from technology companies. It is one thing to say that platforms must mitigate systemic risks; it is another thing to know which mitigation techniques are actually capable of protecting users from skilled adversaries. The project’s findings could influence everything from how the EU assesses the readiness of its member states to how future research funding is allocated. In this sense, FIMI-RESIST is not just a research project, it is a strategic contribution to the democratic resilience of Europe.
Behind the project is a diverse and impressive consortium that reflects the complexity of the challenge itself. The XplaiNLP Research Group at Johannes Gutenberg University Mainz, led by Professor Vera Schmitt, brings together natural language processing, explainable artificial intelligence, and robust information processing. This combination is particularly important because trust in AI-driven tools depends on their ability not only to detect disinformation, but also to explain why they made a particular judgment. The other partners span the European landscape: the Kempelen Institute of Intelligent Technologies in Slovakia, the Centre for Research and Technology Hellas in Greece, the University of Amsterdam in the Netherlands, Logically/INSIKT from the United Kingdom and Spain, Factiverse from Norway, Exorde Labs from France, the Cyfluence Research Center in Germany, Deutsche Welle, and VoxUkraine in Ukraine. Together they bring expertise in fact-checking, data science, network analysis, journalism, and open-source intelligence. The consortium also includes Gretchen AI as an associated partner, and the project has the public support of Alexandra Geese, Member of the European Parliament. The kickoff event at Mainz University brought in voices from the European Commission and civil society, including Saman Nazari from Alliance4Europe and Sandra Fiene from the European Commission, as well as policy and project officers Ilona von Bethlenfalvy and Alberto Domini. It was a reminder that this work requires not only technologists, but also diplomats, policy experts, and engaged citizens.
Over the next 30 months, FIMI-RESIST will compile a wealth of evidence, develop practical recommendations, and build bridges between worlds that too often remain separate: academic research, frontline journalism, government policy, and civil society activism. The deeper aim is to help Europe become more skeptical without becoming cynical, more vigilant without becoming paranoid, and more coordinated without losing the diversity of voices that makes democratic debate meaningful. Disinformation is not a problem that can be solved once and for all; it evolves, adapts, and returns in new forms. But resilience is possible. By learning which countermeasures genuinely work, which ones fail, and how they can be made better together, Europe can improve its ability to protect the information space from those who seek to manipulate it. The work is urgent. Generative AI is making it easier and cheaper to create convincing lies at massive scale, and the same AI models people increasingly rely on for information are themselves vulnerable to manipulation. The tools to fight back exist; they just have not been properly understood, compared, or refined. FIMI-RESIST is the project that aims to change that, not by inventing a perfect shield from scratch, but by strengthening the imperfect shields Europe already has, so that they can be used wisely, fairly, and effectively in the messy, fast-moving, and deeply human world of real information and real influence.