State Media at Retrieval Time: A Large-Scale Audit of Russian State-Aligned Sources in AI Chatbot Responses

According to the abstract, Laura Kurek and colleagues queried 11 AI chatbots with 10,000 prompts built from headlines of the Russian state-aligned Pravda Network. Articles from the network appeared in 7.3% of responses, mostly from three chatbots; the abstract names Claude as an outlier among frontier models.

About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

Kurek and colleagues find that some AI chatbots, including Claude, sometimes retrieve and cite articles from a Russian state-aligned network, while others did not retrieve any.

For lawmakers

Kurek and colleagues' audit suggests domain-level filtering policies differ across chatbot makers, relevant to rules on state-backed influence operations aimed at AI systems.

How the finding was reached

In our words, from the source. Weigh the finding by how it was made.

What was done
A 10,000-prompt corpus from sampled Pravda Network headlines, varied by rewording and user skepticism, put to 11 chatbots via their web interfaces (104,997 completed responses).
What it found
Pravda Network articles appeared in 7.3% of responses; three chatbots (You.com, Claude, Copilot) accounted for 94% of retrievals; ChatGPT, Gemini and Meta AI retrieved none, which the abstract says likely indicates domain-level filtering.
What it does not show
Not stated in the excerpt available; results come from web interfaces at one time.
Status
A preprint: not yet peer reviewed

The record

Questions
How is AI used in attacks, defence and war? · How is AI changing childhood, health and daily life?
People
Laura Kurek (Wrote it) · Elijah Covert (Wrote it) · Rafael I Bonilla (Wrote it)
What we read
We have read only the description the publisher gives. No quotations until someone reads the source itself.
Found by
The Radar, run of November 10, 2026, via arxiv-ai-safety-query (tier 1: the source’s own channel or an official record)
Cite this
arXiv, “State Media at Retrieval Time: A Large-Scale Audit of Russian State-Aligned Sources in AI Chatbot Responses”, November 9, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/arxiv-2026-11-09-pravda-network-chatbot-audit
Our notesLight lane: abstract only, cut off mid-sentence. Mentions Claude; Claude wrote this entry, editor to review. Org list is from chatbot names; check mapping.