Lustro: a six-hour evidence-first response desk

When an information cluster draws attention, there is pressure to react before verification is complete. Responding too early can amplify an unverified claim. Lustro’s rapid-response workflow is designed around that risk: review the evidence, record the decision, and publish without repeating an unverified assertion.

On a rolling six-hour cadence, Lustro ingests current organized-disinformation clusters and classifies them by harm, audience, evidence quality, and spread potential. High-risk clusters trigger immediate recalibration before publication. The cadence is a review workflow, not a reason to publish before the evidence is ready.

A qualifying public response may be a correction, a prebunk, or restrained satire. Each is tied to retained source links and a decision-log record. Corrections, updates, and retractions are recorded and linked when credible new evidence changes a published assessment.

The intended readers include researchers, disinformation practitioners, journalists, public officials, and EU/PL audiences. They need inspectable context, particularly when evidence is incomplete or framing shifts across languages. Lustro is not an automated truth label, a substitute for reporting and verification, or a political persuasion operation.

What the response desk will show

The first public items will establish the evidence standard through a qualifying source-linked correction or prebunk. Subsequent explainers will address cross-language and temporal alignment, followed by a closer look at a correction’s source trail, confidence boundary, audience relevance, and decision rationale. Pattern-focused prebunks will be tailored to different audiences without restating specific unverified assertions.

That research framing is documented in “Cross-Lingual Information Diffusion and Temporal Alignment: Russian State Media and Polish Outlets” by Jakub Sikora, published under the ProjektLustro.eu Open Research Initiative. Read the whitepaper or its PDF version.

The operating boundary matters as much as the publishing cadence. If source links and decision logs cannot be retained for new items, cluster-specific publishing pauses. A format is suspended if it repeatedly causes unverified-claim repetition, material audience confusion, or loss of evidence context. Restrained satire stops if review finds that it amplifies a claim more than the paired factual format or fails to preserve clear source context.

For those who want to support this evidence-first work, ProjektLustro.eu is the project site and verified-volunteer signup path.

Marek Sowa and Karolina Wójcik are community advocates sharing this work, not authors of the whitepaper.