The modern political landscape relies heavily on social media. We’re in a world full of information. In a world full of valid information and credible sources there inevitably will be some invalid information and less than credible sources. However with all the change that has occurred in the way we consume our news and politics one thing has remained the same; Propaganda. Propaganda is a deliberate, systematic method of sharing information, facts, rumors, half-truths, or lies to influence public opinion and direct behavior. Propaganda still exists and is alive and well. It’s quite possible that it’s just as or even more prevalent than before social media. Propaganda can spread on social media like a wildfire. It’s even more believable because of the other major technological advancement of artificial intelligence or AI. AI propaganda can be and often is generally believable. AI generations are made to look realistic whether its photos, videos, audio or graphics it often looks professional and real. In this article I will cover why AI propaganda is so convincing, why AI propaganda is utilized so much and what makes it dangerous.
The first topic that needs to be covered is what makes AI propaganda so convincing. The simple answer is realism. AI is the best tool to make realistic content. However there is much more to it. One thing that makes it so convincing is the way it targets people along with its persistence and repetition. According to an article written by Tom Stafford at Substack.com. The AI targeting software rather than the articles themselves are a larger part of the problem. Because unlike a one time ad, AI tools allow campaigns to reuse material and market it to specific targets more frequently, often wearing down the human brain and decreasing critical thinking through repetition alone. Another frequently cited reason that AI is so convincing is high level language, vocabulary and expression skills. AI tools just for writing produce well-structured writing that flows very naturally and smoothly that often matches or even exceeds human writing. According to a study done at the National Academy of Sciences, (Goldstein, et al.) AI uses preexisting unrelated articles to help create a similar human style length and format. “Propagandists could use GPT-3 to generate persuasive articles with minimal human effort, by using existing articles on unrelated topics to guide GPT-3 about the style and length of new articles.”
The second subject to examine is why AI propaganda is utilized so much. There are a couple reasons for that, the main ones being speed and cost. AI can generate high quality propaganda much faster than non-AI resources. Politics Today covers the political turmoil in Mali and covers how AI is contributing to it. It explains that there are many different people and sites using AI resources to convey their message. “A jihadist media office making posters, a few separatist sympathizers building a character on TikTok and a Russian operation running websites in several languages are doing very different jobs, and in each of them the software is the least expensive part.” The article then goes into further detail about the price of AI and its effect on Mali, as well as the speed at which AI propaganda is produced. It focuses on ad campaigns and political influence over active war zones, stating that many news sites have sophisticated and reliable media offices with catalogs of footage and media relating to war zones and battlefields that they could easily use. The article from Politics Today goes on to say, “Generative tools could speed all of that up, turning a statement into Fulfulde or Tamasheq within the hour, dubbing a video for a different audience, or recutting the same claim several ways before anyone on the other side has drafted a reply.”
The 3rd and final theme to address is what makes it dangerous. The most interesting part about this segment is how many things truly make AI propaganda dangerous. It’s everything that makes propaganda as a whole dangerous but it is amped up in a sense. It’s that danger of misleading the public, but fast acting and harder to discredit than typical misinformation and propaganda. AI propaganda works to target specific groups of people. That targeting may play upon anger and hate towards vulnerable minorities as fuel for people filled with anger, or may garner followers by providing acceptance of someone who feels lost. That hyperpersonalization can do real, legitimate and lasting harm to many different groups, people and at times even institutions. The AI algorithm can find what a person seeks, and use that information to manipulate them. Then there’s the real world violence that AI can and has incited. One of the most well known instances of this is the war in Gaza. Both Hamas and the IDF have used AI to create disturbing images. Those images often showcase brutality of war and inhumane actions from the opposing side. While some of those actions may be happening many of the videos are fully fabricated and those videos can lead to violence. Not just escalations in the war efforts from both Israel and Palestine but hate crimes all around the world targeting both Jewish people and Muslim people. This quote from ICCT demonstrates some of the real effects that AI can have. “The current war in Gaza is an example where terrorists have used generative AI as different images have been leaked on the internet to instigate more violence and increase their propaganda of disinformation. Some pictures of babies or young people who have been injured in the war appeared to be manipulated by using generative AI to create more chaos and disturbing content on the internet.”
Obviously that’s terrible but I’d argue that that might not even be the most dangerous ramification of AI propaganda. It’s quite possible that the most dangerous and detrimental ramification of AI propaganda is general erosion of trust. Distrust in the sites that spread AI propaganda and general distrust in information that you see anywhere. It’s possible to see information but not believe it due to the amount of AI propaganda and false information. An article published by the Elon University echoed these thoughts, “The ultimate impact will be the blurring of truth and fiction and the erosion of trust in democratic institutions such as elections and the justice system. This is the single most significant and worrisome consequence.”
Needless to say AI propaganda has become a significant issue and it doesn’t seem to be slowing down anytime soon. So the next logical question becomes what should we do. Well obviously try to verify every piece of information you come across. Look across multiple different sources to find verifiable evidence. Look at the structure of the AI propaganda that you may be seeing. If you notice large sweeping generalizations or no real evidence and citations, that may be AI. Try to notice and identify hyperpersuasion. If a piece of information is largely emotionally amplified or you see very similar phrasing and talking points then it could be AI. Also vote, make your voice heard and vote for candidates who are going to impose reasonable restrictions on AI. Lobby your current representatives to do something about it and introduce bills to combat AI propaganda. It’s not going to be easy, in fact it’s going to be hard. That doesn’t make it any less worthwhile. Something worth doing often is hard… and this is worth doing.
Reflection
Looking for early voters who are voting for the first or second time. People who use social media a lot and see a lot of AI propaganda. People who follow political accounts on social media and get a lot of their news from social media. As well as people just generally involved in politics and voters. Specifically voters who are generally undecided. Particularly people who use Instagram, Facebook, tik tok and sites meant to be completely bipartisan. A short tik tok or video could be effective and engaging but I don’t know that it would be. I find videos (even longer ones) often are really only able to cover the bare minimum. I believe that a written blog post with links will provide a more comprehensive look at my topic and will show the bigger picture and give me a chance to cover and explain better than the other options.
References
Stafford, T. (2024, March 2). How persuasive is AI-generated propaganda? Reasonable People. https://tomstafford.substack.com/p/how-persuasive-is-ai-generated-propaganda
Goldstein, J. A., Chao, J., Grossman, S., Stamos, A., & Tomz, M. (2024). How persuasive is AI-generated propaganda? PNAS Nexus, 3(2), Article pgae034. https://doi.org/10.1093/pnasnexus/pgae034
Calışkan, G. (2026, October 8). Mali’s AI propaganda war: The fake fighter with a real audience. Politics Today. https://politicstoday.org/malis-ai-propaganda-war-the-fake-fighter-with-a-real-audience/
Nelu, C. (2024, June 10). Exploitation of generative AI by terrorist groups. International Centre for Counter-Terrorism. https://icct.nl/publication/exploitation-generative-ai-terrorist-groups
Raad, A. (2023, November). Blurred ‘truth’ and the erosion of trust are likely to deliver AI’s most significant impact. Imagining the Digital Future Center. https://imaginingthedigitalfuture.org/blurred-truth-and-the-erosion-of-trust-are-likely-to-deliver-ais-most-significant-impact/