Propaganda in the Age of Artificial Intelligence

This article is a column; all content and opinions are those of the author and do not necessarily reflect the editors’ stance. The letsdatascience.com edition of 8 May 2026 reports on the Vietnamese Government’s plan to massively harness artificial intelligence (AI) together with its experts. The utilisation is framed as the country’s ‘positive’ communication, to create an ‘ideological resilience’ for all its citizens. Immunity to harmful, toxic, and false information. This positive communication is, in effect, propaganda. The full briefing published as “Vietnam Recruits Influencers and AI Experts for Propaganda” is a quotation from a Reuters report. It states that the ruling party’s propaganda committee in Vietnam up to 2030 plans to build a network: 1,000 influencers and 5,000 AI experts. In the April-drafted document, it is also stated: when the positive communications have been rolled out to the end of the decade, at least 80 per cent of online posts—presented in Vietnamese—will be ‘positive’. This is because AI, within 24 hours, will remove up to 90 per cent of material that violates party guidelines. Interactions are also conducted in formats such as podcasts, short videos, and various forms of posts for a young audience. The execution of these interactions is supported by AI-based devices developed by Vietnamese technology companies. Their aim is to shape, control, and even alter the attitudes of other groups, using instruments of communication. The result is propaganda-targeted behaviour—whether among citizens or political constituents—that aligns with the propaganda producer’s wishes. Meanwhile, Harold D. Laswell, 1927, in “The Theory of Political Propaganda” described propaganda as the management of collective attitudes through the manipulation of key symbols. There are many other definitions cited by scholars on propaganda. Yet all converge on manufacturing audience behaviour, caused by one-sided message delivery. The purpose of sending the message is to create the reaction the message producer desires rather than the truth of the information. And when truth is not a central part of the message, propaganda becomes disinformation, misinformation, or malinformation that is widely disseminated. Today, civilisation calls it hoaxes. What are the consequences of state-sponsored propaganda, especially when leveraging AI? This research seeks to unravel the consequences of AI adoption in state-sponsored propaganda, focusing on Russia, using quasi-experimental, text-based, and survey methods. It found that state-sponsored propaganda through the adoption of generative AI technology enables the websites used to strengthen and increase disinformation production. Specifically, first, Gen-AI tools facilitate larger-scale disinformation production than relying on human labour alone. Second, the use of Gen-AI is closely linked to shifts in both the volume and scope of posts published; volume is higher and reach broader. And third, this finding reveals the power of AI-powered propaganda—based on a survey experiment—that persuasiveness of AI-assisted articles endures into the post-adoption period.