War Room co-host Natalie Winters talked about China using AI to build models of US voters. She explained that it involved them looking at things like why they vote a certain way, understanding President Trump’s base, and even how they view China.
Winters explained that the AI models include mock US elections to determine potential outcomes.
“Walk me through your CCP story in this regards ma’am,” Steve Bannon asked.
“I think the AI issue really is the convergence right, of the decade. The CCP has spent deploying assets compromising people, setting up these information networks. You are seeing all those people be deployed right now,” Winters said.
“That is where the United States still retains most of our military edge over the Chinese Communist Party. Right, we are very advanced in comparison to where they are, so they want to basically square that gap, right, but AI is going to be the way that wars will be fought in the future, right, so that’s why you have got to look at this,” Winters explained.
“This is sort of an edge case application of it, though it is very mainstream in the Chinese academic and military research community, but this dovetails exactly with what President Trump was talking about in the Oval Office, what was it about a month ago, about the voter rolls, election influence, election intimidation,” Winters said.
“So, basically, you have two kinds of categories of the research that the PRC is conducting. One, is they are targeting and trying to understand the American voting bloc. The demographics, particularly trying to understand President Trump’s base, which I think is quite interesting, the kind of populist right as they call it, the MAGA audience. But also just more broadly what makes Americans tick, what makes them vote for certain candidates. How can you get turnout. What are their perspectives on China,” Winters commented.
“Then you have a more broader simulation tier which they are running, which is actually running mock US elections,” Winters said.
“If you wanted to influence American elections, this is how you would be doing it. Again, these are studies that are published, but not necessarily for public consumption,” Winters continued.
“They are basically testing which messages will move voters, predicting how people will react. They are mass producing essentially these bot farms,” Winters said.
“For example, during the 2024 race, researchers at Beijing University collected 1,000 Trump, Harris posts, and over 25,000 replies to them. So, they trained AI off of those models, the LLMs. They trained them to be able to respond to future posts from President Trump or Kamala Harris, or whoever, and be able to basically mock or simulate a reply from an American to most likely attack President Trump,” Winters explained.
“They say in their reporting that it can ‘help shape public opinion in the United States,’” Winters said.
“It’s about manipulating voters, manipulating their beliefs, their ability to turnout,” Winters said.
Watch:
Ad block users: Some site features may not work correctly while an ad blocker is enabled, because they break scripts and content this website depends on. If you can’t see comments below, for example, please disable your ad blocker.
