Coordinated networks linked to Pakistani military, PR firms and state actors have spread anti-India content using fake personas, AI tools and amplification tactics.
A 2025 study found over 24,600 such posts with 300 million views.
The networks operate as influence supply chains combining state employees, paid actors, volunteers and algorithms—not simple botnets—designed to normalize hostility and create crisis confusion.
What happened
Meta has linked Pakistan-origin networks to Pakistani military communications staff and a Pakistan-based PR firm AlphaPro; Google removed 447 Pakistan-linked YouTube channels between late 2025 and mid-2026 that supported Pakistan and criticized India; Stanford researchers documented a Pakistan-based mass-reporting network using a Chrome extension; and OpenAI disrupted an Israeli political campaign firm (STOIC) using AI-generated material in India's 2024 election.
Why it matters
The research reveals coordinated inauthentic behavior operating as an influence supply chain—combining state employees, PR firms, fake media brands, AI tools, paid actors, volunteers and recommendation algorithms—rather than a single botnet. A 2025 analysis counted more than 24,600 anti-Indian posts from nearly 14,000 authors with over 300 million reported views, with much activity amplified by Western nativist and immigration politics. The goal is to normalize contempt, exhaust targeted communities and create confusion during crises by moving false claims through fake identities into verified accounts and news channels.
What to watch
During the May 2025 India–Pakistan crisis, researchers collected roughly 1,200 misleading posts; only 73 of 437 posts sampled from X had Community Notes at the time, showing a detection gap. Visual manipulation is routine—cases include a 2023 naval-drill image falsely presented as a Karachi attack, Gaza footage relabeled as Operation Sindoor, and AI-generated combat narratives. The research distinguishes platform-verified findings (Meta and Google takedowns) from researcher-based indicators and unverified allegations to avoid false attribution.
Ask the AI about this article →
The research frames identity-based hatred as a low-cost cyber operation targeting human perception rather than computer infrastructure. Unlike traditional cybersecurity attacks with defined endpoints, influence networks operate across a supply chain where state employees, PR firms, fake media outlets, paid actors, volunteers, influencers and algorithms all participate in coordinated inauthentic behavior. The distinction between fully automated botnets and human-operated networks matters: a hundred humans posing as independent users can be more persuasive than obvious automation, and bot-detection tools often disappoint because they cannot conclusively distinguish between harmless news bots, very active humans and deliberate influence operators posing as normal users.
The strongest public findings come from platform attribution based on internal login and infrastructure data—Meta's connections to ISPR and AlphaPro employees, Google's removal of 447 channels, and OpenAI's disruption of STOIC within 24 hours. However, the research methodology distinguishes three confidence levels: platform-verified takedowns, researcher-based indicators (like BOOM's detection of troll-farm signs without identifying the operator), and unverified allegations. The goal of these networks is not necessarily to make everyone believe one lie, but to make contempt feel normal, exhaust targeted communities, create confusion during crises and push regular users into amplifying content for free through their own social networks.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
