evidoria

← Back to browse

Good practice Imported

Debunk.eu — Lithuania's AI-Assisted Disinformation Detection Network

Lithuania · Vilnius · See the Lithuania profile

Since 2018, Lithuania's Debunk.eu has used AI to scan roughly 30,000 articles a day across 2,500+ web domains in 26 languages, triaging hits to about 5,000 vetted volunteers, backed by Lithuania's Defence and Foreign Ministries — though no independent impact evaluation was found.

Debunk.eu — Lithuania's AI-Assisted Disinformation Detection Network

Details

Promoter
Debunk.eu (Debunk Community)
Period
2018–present
Keywords
civic tech, disinformation monitoring, media literacy, national security communications

Description

Debunk.eu (originally the Demaskuok project) was founded in Vilnius in 2018 by Viktoras Daukšas, combining an AI-based content-scanning tool with a network of vetted volunteer analysts known as "elves" to identify and respond to disinformation targeting Lithuania and the wider Baltic region.
According to Debunk.eu's own account and an independent Wikipedia summary of third-party reporting, the platform's AI algorithms scan roughly 30,000 online articles per day across more than 2,500 web domains in 26 languages, searching for keyword patterns and over 600 known propaganda and disinformation narratives. Machine-flagged content is triaged by a volunteer network of around 5,000 trained "elves," with about 50 collaborating directly with the core analysis team, enabling identification and rebuttal of suspected disinformation within roughly two hours. The project has received funding from Lithuania's Ministries of Foreign Affairs and Defence, the German federal government, the German Marshall Fund, the UK Foreign Office, and a €315,000 Google Digital News Innovation grant, and coordinates with state communications officials in real time.
No independent, peer-reviewed evaluation of the platform's actual effect on disinformation belief or spread was found; the available figures describe monitoring throughput and volunteer scale rather than measured downstream impact, and the organisation's own materials do not disclose model accuracy, false-positive rates, or an external audit of its classification method.

Read the full analysis: https://debunk.eu/about-debunk/

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful