On September 10, 2026, Euronews published a fact-check of a claim by Florian Philippot, president of France's Patriots party. He said the European Democracy Shield "could lead to the annulment of the electoral process in all EU countries, as well as France's 2027 presidential election," as happened in Romania. Euronews found no such mechanism. It noted that "the conduct of elections is a competence of individual Member States." The claim is wrong. But the episode shows how much of the public misunderstands who decides what comes down online in the EU, and that confusion is itself a policy problem.
The strongest case for the Shield
The case for the Shield should be stated fairly. Coordinated foreign influence operations are real. Romania's Constitutional Court annulled its 2024 presidential first round after intelligence assessments alleged Russian-linked manipulation, including on TikTok, according to Euronews. Fact-checking is also a comparatively speech-friendly tool. It adds information rather than deleting it. The Commission's Communication on the European Democracy Shield, published on November 12, 2025, frames its approach as strengthening democratic resilience while, in its words about the Code of Conduct on Disinformation, "fully upholding the freedom of expression." If you are worried about disinformation, funding independent fact-checkers beats most of the alternatives.
What the Shield does and does not do
The Commission announced a €5 million grant for a European fact-checking network on March 31, 2026. It is led by the European Fact-Checking Standards Network (EFCSN) with seven partner organisations. The funding covers fact-checking capacity in all EU languages, a protection scheme for fact-checkers (legal, cybersecurity and psychological support), and an independent repository of fact-checks. The Commission's announcement contains no takedown power.
The Euronews analysis matches this. Fact-checkers can identify and flag claims. They cannot remove content. Election administration, including annulment, stays with member states and their courts. Romania's decision was made by its own Constitutional Court under domestic law. Brussels did not make it.
Where the real line sits: the DSA
Removal authority in EU law comes from the Digital Services Act (Regulation (EU) 2022/2065), not from the Shield. Three features matter:
- Illegal content: Article 16 requires hosting providers to run notice-and-action mechanisms for allegedly illegal content. Article 17 requires a statement of reasons to affected users. Article 22 gives prioritised treatment to notices from certified "trusted flaggers."
- No general monitoring: Article 8 bars general monitoring obligations.
- Systemic risk: Articles 34 and 35 require very large platforms to assess and mitigate risks to civic discourse and electoral processes. This is the DSA's most flexible lever, and it does not name specific content to be removed.
Legal but false claims occupy the gap. The DSA does not make ordinary misinformation illegal. The voluntary Code of Conduct on Disinformation, tied to the DSA, is where platforms make commitments on demonetisation, labelling and fact-check integration. Tech Policy Press reports that the Commission has leaned on this voluntary approach and "light-touch" dialogue rather than formal proceedings, even after alleged interference in German and Romanian elections, and that several major participants have left the Code.
The proportionality concern that deserves attention
Philippot's claim is false, but a narrower worry is legitimate. Flagging is not removal, yet flags from a publicly funded network feed platforms that face systemic-risk duties under Articles 34 and 35. A platform worried about regulatory exposure may over-remove to be safe. Nothing in the text orders that outcome, but the incentive exists, and it is the kind of soft pressure that speech advocates should watch. The Commission's own Communication says the Code helps mitigate risks while upholding free expression. Whether that holds in practice is an empirical question, and current oversight gives us little data to answer it.
Three safeguards would keep flagging from sliding into takedown by proxy:
- Publish platform actions taken after fact-checker flags, with outcomes, so researchers can measure over-removal.
- Apply Article 17 statements of reasons and appeal routes to any demotion or removal that follows a flag, not only to formally illegal content.
- Keep fact-checker independence real. The repository of fact-checks should be open to challenge, with a correction process.
The takeaway for policy debates
The misinformation debate in Europe suffers from two opposite errors. One treats fact-checking funding as a censorship engine that can cancel elections. The other treats any scrutiny of the Shield as a disinformation talking point. The evidence supports neither. The Shield adds capacity for flagging and coordination. Removal authority still runs through the DSA's illegal-content rules and platforms' own terms of service, and elections through national courts.
That is a defensible structure, and it is better than a legal ban on "false information," which no EU instrument imposes. The test of the next year is whether platforms treat flags as information or as instructions. Transparency on that point, not rhetoric from either side, will settle it.