Chinese state researchers are using LLMs and AI models to map how to target susceptible Americans and influence their votes using public source information.

Posted by Bannon`s War Room, Aug 17, 2026Checked Nov 2, 2026

Polygraph verdictTrue

Sources: 1 contradict4 support
Said: 1 time·
Confidence: 60%

Bannon`s War RoomAug 17, 2026

Sources on the same subject support the core assertion that China-linked state actors use AI to profile Americans and run influence operations, including fake American personas and profiling of US politicians (GoLaxy, Spamouflage).

This claim is about the present, so it is judged as of the day it was checked.

What was said

… So, this is, again, like, Swawal, it's one small example, part of a broader thing. But the bigger thing here, and if we have time, we'll get into it after the break, is how this all converges, which is forthcoming exclusive: that Chinese state researchers are mapping how to target, quote, susceptible Americans and influence their votes. They're using these LLMs and these AI models to get all this public source information about Americans, particularly from social media posts and the voter files, all the publicly available information, creating mock Americans, and then figuring out the best way to literally change their votes. We'll get into the quotes. This is actual Chinese state-funded research, and improve their opinions of China. This is, like, how it all converges and what the point of all this comes to. …

Bannon`s War Room · Aug 17, 2026· from the transcriptPost →

What we found

Supports. A Substack review published 2026-08-17 reports Chinese state-linked studies that classify Americans by 'degree of susceptibility' and propose messages to move their voting intentions, though it is an unrated single outlet.[4]

Supports. Vanderbilt researchers found documents showing the Chinese firm GoLaxy used AI to generate social media profiles and map existing ones to influence US politicians and influencers.[5]

Supports. Chinese-linked firms reportedly compiled profiles on at least 117 members of Congress and more than 2,000 American political figures, per Vanderbilt research cited by CNN.[3]

Supports. AP and Graphika traced a Spamouflage network that used fake, likely AI-generated American personas to mimic Americans and sway US political debate.[1]

Contradicts. A Microsoft-based report summarized by the MIT AI Incident database concludes that foreign influence campaigns, including China-linked ones, failed to achieve measurable results in the 2024 US elections.[2]

Context. The sources do not address the claim's reference to voter files as a data source, and the claim's wording about 'changing votes' describes an aim rather than a documented outcome.[2][4]

Sources

5 sources: 1 fact-checker, 1 primary, 1 tier 2, 2 not rated

1China-linked 'Spamouflage' network mimics Americans ...apnews.com · fact-checker · cited for: supports
2China Reportedly Intensifying AI to Spread… – AI Incident 663airisk.mit.edu · primary source · cited for: contradicts, context
3Secret US cyber operations shielded 2024 election from foreign trolls, but now the Trump admin has gutted protections | CNN Politicscnn.com · tier 2 · cited for: supports
4Chinese State Researchers Mapped How To Target ...nataliegwinters.substack.com · not rated · cited for: supports, context
5Chinese biz using AI to influence US politicians - The Registertheregister.com · not rated · cited for: supports

Where this claim was made

Said in 1 published post

Bannon`s War Room · Aug 17, 2026Post →

How Polygraph checks a claim

What was said

What we found

Sources

Where this claim was made

Related claims

The Readout