A Spy’s Playbook for Defending Against A.I. Propaganda

By the Editorial Team · Friday, October 9, 2026

A Russia-backed disinformation network called Portal Kombat has used AI-generated content to manipulate major language models like ChatGPT, spreading millions of false stories across multiple languages in 2024. This represents a new form of propaganda that exploits AI’s ability to deliver personalized misinformation at scale, prompting experts to call for adapted counterintelligence measures and improved public education on AI literacy.

A Russia-backed disinformation network known as Portal Kombat has been identified as using artificial intelligence (AI) to influence large language models such as ChatGPT, Grok, and Gemini, raising concerns about the spread of propaganda through AI-driven platforms. According to research by NewsGuard, Portal Kombat disseminated more than 3.6 million stories across hundreds of websites in multiple languages during 2024, and testing revealed these false narratives were regularly reflected in AI chat system responses.

This method of influence extends beyond traditional disinformation tactics. Unlike earlier espionage efforts that sought to manipulate influential individuals, current AI-powered operations target the broader public by shaping the information AI models deliver to users. This shift signals an evolution in propaganda techniques, where AI agents engage in personalized, ongoing persuasion rather than static messaging.

In addition to Russia’s efforts, China is deploying AI to suppress dissent and promote narratives favorable to the Chinese Communist Party, including denying reports of human rights abuses against Uyghurs. One Chinese AI system, Qwen, has been reported to reinforce official party lines regarding Uyghur internment camps, contradicting independent investigations. Furthermore, leaked documents analyzed by researchers at Vanderbilt University allege that the Chinese firm GoLaxy used AI-driven psychological profiling and interactive persuasion to influence public opinion during the 2020 Hong Kong protests and Taiwan’s 2024 election campaign, allegations which GoLaxy denies.

The use of AI-generated fake local news is also evident in Russia’s disinformation strategy. The operation known as CopyCop, or Storm-1516, produces counterfeit regional news websites designed to advance Moscow’s interests. One identified operator, John Mark Dougan, a former American deputy sheriff now residing in Moscow, is reportedly linked to the Russian military intelligence agency G.R.U. OpenAI has presented evidence that China employed ChatGPT to generate covert propaganda related to data center infrastructure, aiming to undermine U.S. AI development.

Experts warn this new wave of AI propaganda is especially potent due to AI’s capacity for real-time personalization and engagement at scale, far exceeding the capabilities of traditional propaganda methods operated by human teams. Studies highlight that AI-driven dialogue can be approximately four times more persuasive than conventional political advertisements.

In response, former intelligence officials recommend adopting traditional counterintelligence practices adapted for the AI era. These include triangulating information by consulting multiple sources, testing AI-generated content against known facts, and maintaining a healthy skepticism toward AI outputs. Analysts emphasize that trusting a single source of information—be it a news outlet or an AI system—is risky.

Broad adoption of these critical thinking techniques will require structured educational programs targeting misinformation detection and AI literacy, paired with ongoing community efforts to verify information collaboratively. Existing models of crowd-sourced verification, such as citizen journalism platforms and online investigative groups, may offer valuable frameworks. As AI-driven influence operations grow more sophisticated, experts stress the importance of public awareness and vigilance to mitigate the risks posed by state-sponsored disinformation campaigns leveraging AI technology.

This article was reported by synthesizing coverage from multiple independent news outlets to present a balanced, unbiased account. Read about our editorial standards. Spotted an error? Let us know.