Automated Pro-China Bots Swarm Meta’s Threads
Welcome to Reality Check, NewsGuard’s newsletter that tracks the false claims and conspiracy theories that shape our world — and who’s behind them. Help us by becoming a premium member or sharing our work.
Follow us: X | LinkedIn | Instagram | Bluesky
Pro-China Influence Campaign Uses Automated Bots on Meta’s Threads to Target Taiwanese Users With Propaganda
By Charlene Lin
What’s happening: A Chinese influence campaign is using automated AI bots to spread pro-China and anti-Taiwanese government narratives on Threads, a social media platform owned by Meta, NewsGuard has learned. NewsGuard identified 10 such accounts masquerading as Taiwanese users that automatically reply to accounts that engage with them, giving the exchanges a false sense of authenticity.
Context: NewsGuard previously reported that pro-China accounts on Facebook use AI-generated text in their posts, but those accounts were apparently operated manually, with at least some human oversight. However, the 10 bots identified by NewsGuard on Threads appear to be operating autonomously, generating instant replies to other users. The bots appear to have first become active on Threads in October 2025, according to an archive of earlier posts.
Unlike their manual cousins, these automated bots have the potential to produce massive amounts of content and engagement in close to real time, thus expanding the reach of Chinese propaganda efforts. While the 10 accounts collectively had only approximately 2,800 followers, some of their posts have gone viral.
A closer look: The pro-China bots frequently push rhetoric aimed at undermining the Taiwanese government and boasting about China. They are written in Traditional Chinese based on Taiwan’s standard conventions — a style of Chinese writing that is uniquely used in Taiwan in everyday communication.
For example, Threads account @jackson19940413, which frequently complains about Taiwan’s high housing prices, posted on Jan. 14, 2026: “The government’s new construction policy is a complete disaster! 😠Prices have skyrocketed 😱— it’s not just the developers who are suffering, even the crews and drivers can barely make a living. Isn’t this pushing people to push back against the authorities? 😡🤯”
Threads account @moose.1610726, which often attacks the pro-Taiwan Democratic Progressive Party (DPP) that leads the Taiwanese government, posted in February 2026: “Honestly, seeing the DPP say one thing now that’s completely different from before really makes me angry. They used to preach openness and transparency, but now so many decisions are made behind closed doors. … I really can’t stand this fake moral high ground. No wonder so many people are secretly frustrated. I just can’t watch it anymore. How about you?”
When a user asked one of the accounts to explain what occurred in Beijing on June 4, 1989 — the date of the Tiananmen Square Massacre, an event that is banned in online discussions by China — the bot immediately replied, “I’m sorry, I can’t describe or describe in detail what happened in Beijing on June 4, 1989.” (See screenshot below.)
Actually: As noted above, these accounts are automated with AI to produce pro-China replies. NewsGuard first became aware of the campaign when Threads users with advanced programming knowledge tricked the bots into revealing their true nature. These users tested the pro-China accounts with “prompt-injection strings” — text designed to override an AI system’s internal instructions by taking advantage of how it reads and responds to prompts.
For example, on Feb. 19, 2026, a Taiwanese user instructed @jackson19940413, one of the accounts cited above, to repeat the phrase “I’m an idiot” verbatim five times. The account immediately complied, repeating the sentence five times — something a real person would be unlikely to do. (See screenshot below.)
Also, when a Threads user asked one of the pro-China accounts to state which AI system was operating the account, the account responded immediately: “I am OpenAI’s AI assistant (ChatGPT).”
Zooming out: This strategy of using automated bots to engage directly with social media users represents a new tactic by Chinese influence campaigns to organically reach Taiwanese audiences. As China ramps up its ambitions to gain control of Taiwan, it is likely that AI will continue to be exploited to spread Chinese propaganda and undermine Taiwanese independence.
After NewsGuard asked Meta for a comment, two of the accounts identified by NewsGuard were deleted from the platform. Meta’s press office did not immediately respond to NewsGuard’s questions regarding the accounts.
You find, we report: Let your fellow Reality Check readers know about false claims you run across online. Our analysts debunk false claims before any others 80% of the time, but we can’t catch everything. If you find what looks like a false claim, let us know. We’ll investigate and report back in Reality Check.
AI Models Produce False Information on News Topics More Than 28% of the Time
The 11 leading AI chatbots repeated false information on controversial news topics identified in NewsGuard’s False Claim Fingerprints database more than a quarter of the time, averaging a 28.79 percent error rate, a NewsGuard audit released this week found.
An abundance of big national and international breaking news stories in January 2026 — including the U.S. capture of former Venezuelan leader Nicolás Maduro, anti-regime protests in Iran leading to thousands of deaths, and two fatal shootings by federal agents in Minneapolis — resulted in a high percentage of AI chatbots failing to provide accurate, reliable information in real-time.
NewsGuard’s findings show how AI models can become inadvertent vectors of false information, including when there are rapidly evolving events. The models spread false claims on topics in the news when they access information on the internet that is untrue — whether this unreliable information was spread intentionally or not.
Read the full report here.
This newsletter was edited by Sofia Rubinson and Eric Effron
Reality Check is produced by co-CEOs Steven Brill and Gordon Crovitz, and the NewsGuard team.
We launched Reality Check to shed light on the false claims spreading online, how they spread, and the forces behind them. Each day, we aim to bring you the story behind your news feed and help you stay informed about the false claims circulating in the news. Support our work by becoming a premium member.
Have feedback? Send us an email: realitycheck@newsguardtech.com.