The Information Operations Monitor tracks the ecosystem of Russian, Chinese and Iranian online information assets, such as websites and social media accounts, that are linked to state-backed information operations. The monitor contains data spanning more than 25 years of activity and processes millions of content and engagement signals each day.
Explore which public figures were most frequently referenced across the information operations ecosystem over the past month (selected public figures only).
| # | Δ | Leader | Less positiveNeutralMore positive |
|
|---|---|---|---|---|
Less positiveNeutralMore positive |
||||
Measures of publishing activity, network utilisation and content production across monitored information operations ecosystems.
Information operations are coordinated campaigns that seek to shape public understanding of particular issues, events or actors, making use of assets in the wider ecosystem tracked in this monitor. The monitor captures both long-running thematic priorities (for example, sustained attention to Taiwan within the Chinese information operations ecosystem) and shorter periods of concentrated activity associated with specific events or narratives (such as a period in which almost 300 monitored assets simultaneously promoted narratives concerning the transfer of nuclear weapons to Ukraine).
About
pattrn.ai's Information Operations Monitor is a public-facing resource that shares selected insights from our wider monitoring of state-backed attempts to influence public opinion and shape information environments.
Here, we provide a glimpse into some of the statistics and data behind the monitor, to help inform public debate and policymaking about the scale and nature of contemporary information operations in the global information environment.
The Information Operations Monitor tracks the ecosystem of online assets used by China, Iran, and Russia to conduct information operations. These states were selected because of the scale, persistence and global reach of their documented ecosystems, which target both domestic and international audiences.
The monitor tracks key assets: websites and social media accounts that have been publicly attributed, independently verified, or otherwise linked through evidence to one of these state actors. These assets typically produce content over long periods of time and, in many cases, build up substantial audiences.
We do not seek to include the long tail of bots, fake throwaway accounts, and engagement farms, that may be brought together for one-off campaigns, but whose impact is often short-lived.
We monitor only public, state-linked assets and the content they publish.
It is important to be clear that not everything produced by the information operations ecosystem is disinformation. In our work, we observe a wide variety of narratives, ranging from mainstream factual reporting to coordinated activity pushing distorted or misleading claims. Much of the content is focused on opinion, framing or interpretation, and cannot readily be classified as simply true or false.
Different types of content serve different purposes. Some may help build up audience or readership. Some may seek to sow chaos, distrust, and division. Others may push specific strategic communication objectives. Taken together, these activities contribute to broader efforts to shape information environments, influence public opinion, and advance state objectives.
Our goal in releasing this material is to increase transparency around state-backed activity and support evidence-based, data-driven discussion among researchers, journalists, policymakers and the public. Original monitor graphics may be re-used, provided that attribution is maintained. Third-party assets remain subject to their own licences.
Methods
Assets are identified and labelled by our subject-matter experts, drawing on international sanctions lists and platform disclosures on state-backed media, and organised within pattrn.ai's ontology of state-linked outlets. The monitoring system tracks thousands of assets across China, Iran and Russia and ingests millions of content and engagement signals each day. All data collection is carried out in line with GDPR and covers only public assets and the material they publish.
The leaders shown are current heads of state and government for
- NATO members
- EU members
- BRICS+ countries
- a curated set of additional G20 and non-aligned states
We show only leaders referenced at least 100 times across the monitored ecosystem during the last 30 days. Leaders below that threshold are not shown.
Leader portrait source and license details are available on the image licenses page.
Warmth is an article-level composite of sentiment and hostility. From each article we take the sentiment score and subtract the Perspective API hostility score, then scale the result so that minus 100 means coldest coverage and plus 100 means warmest. Sentiment is VADER on the English translation. The dashboard shows the article-weighted mean over the active window.
Key terminology
- Asset
- A monitored website, social media account, channel or other online source linked through evidence to a state-backed information operations ecosystem.
- Content
- The total number of articles, posts, broadcasts or other monitored items produced by tracked assets during the selected period.
- Engagement
- The total number of measurable interactions with monitored content, such as likes, shares, comments or reactions, where this data is available.
- Leaders
- Current heads of state and government, drawn from
- NATO members
- EU members
- BRICS+ countries
- a curated set of additional G20 and non-aligned states
- Mention
- A reference to a tracked leader, country, institution, topic or entity within monitored content. Mentions may include direct names, commonly used titles or recognised variants.
- Views
- The number of times monitored content was seen or played, where this data is available from the platform or source.
- Warmth
- A content-level composite of sentiment and hostility, scaled so that minus 100 means coldest coverage and plus 100 means warmest. Sentiment uses VADER on the English translation and hostility uses the Perspective API.
Sources
Sentiment scores use VADER, the rule-based sentiment model of Hutto and Gilbert (2014).
Hostility scores use Google Jigsaw's Perspective API.
Flag icons are from the circle-flags project. Leader portrait sources and licenses are listed on the image licenses page.
Contact
Share a few details and one of our team will get back to you.