A Russia-backed disinformation network known as Portal Kombat has been exploiting artificial intelligence (AI) systems to spread propaganda and influence public opinion, according to recent findings. Operating hundreds of websites and distributing content in dozens of languages, this group aims to manipulate AI-driven large language models such as ChatGPT, Grok, and Gemini to disseminate false narratives to millions of users.
In 2024, Portal Kombat reportedly pushed over 3.6 million stories online, leveraging the expansive reach of AI platforms. Researchers at NewsGuard, an organization monitoring disinformation, tested various AI chatbots and discovered the network’s propaganda frequently surfaced in responses provided by these systems.
Beyond Russia, other state actors are similarly employing AI to further political agendas. China, for example, uses AI platforms to promote misleading narratives about the treatment of Uyghurs in Xinjiang and to suppress dissent related to the Chinese Communist Party. In another instance, China’s perceived efforts to interfere with Taiwan’s 2024 election involved an alleged AI-driven persuasion campaign by the company GoLaxy, which mined social media data to craft personalized content aimed at swaying political views. GoLaxy denies these allegations.
In addition to written disinformation, AI-generated fake audio and video content has become a tool for covert influence. One notable example includes CopyCop, also known as Storm-1516, which produces realistic local news articles favorable to Russian interests under the guise of credible U.S. news outlets such as "The Boston Times" or sites mimicking established newspapers. French authorities have linked this operation and one of its operators, John Mark Dougan—a former Palm Beach deputy sheriff now residing in Moscow—to the Russian military intelligence agency G.R.U.
OpenAI has also reported China’s use of ChatGPT to craft covert propaganda messages linked to campaigns critical of U.S. AI development, such as the so-called “Data Center Bandwagon” effort that involved posting misleading social media content about electricity costs.
Experts warn that the primary threat from AI-driven disinformation is not merely the technology’s capacity to fabricate falsehoods but its unprecedented ability to persuade through personalized, real-time interactions. Unlike traditional propaganda, AI can tailor its messaging to individual users, identifying their beliefs and desires to maximize influence rapidly and at minimal expense.
To counter these tactics, intelligence specialists emphasize adopting verification techniques akin to those used by CIA officers during the Cold War. These methods include triangulating information across multiple sources, operationally testing AI responses against verifiable facts, and maintaining a healthy skepticism toward any single source of information. Users are encouraged to cross-check AI answers with independent news outlets and primary data to mitigate the risk of deception.
The personalization and adaptability of AI tools represent a significant challenge in distinguishing fact from manufactured narratives. However, analysts suggest that by applying structured verification approaches and fostering critical thinking, individuals can better navigate the complex information environment shaped by adversarial AI operations.
At a time when AI-driven disinformation campaigns threaten democratic processes and social cohesion, these countermeasures offer a path toward greater resilience against manipulation and misinformation.