Artificial intelligence has become a key factor in the spread of misinformation about the war in Ukraine, particularly through the creation of fake images and the manipulation of chatbots like ChatGPT, which can disseminate false information aimed at influencing perceptions of the conflict, which has now lasted four years. Since the onset of the war, Ukraine has been the epicenter of a wave of hoaxes and misleading narratives that have crossed borders, but the arrival of AI has radically transformed the scale and reach of propaganda, according to Olga Petriv, an AI expert and lawyer at the Center for Democracy and the Rule of Law in Ukraine. Petriv notes that misinformation has evolved from the manual fabrication of fake news to large-scale automated campaigns, where one of the main objectives is to manipulate the responses of conversational chatbots. One of the most concerning techniques is called LLM Grooming, which involves saturating the internet with misinformation so that language models pick it up and reproduce it in their responses. “By flooding the web with millions of AI-generated articles, propaganda networks contaminate the information environment, causing generative AI systems to repeat these falsehoods as if they were facts,” explains Petriv. An example is the Pravda network, a conglomerate of websites present in dozens of countries, which publishes thousands of monthly articles aligned with the pro-Russian narrative. Research indicates that conventional media and blogs link to this content, increasing its visibility and credibility in both search engines and chatbots. The development of AI image generation tools has also facilitated the creation of visual content that is almost indistinguishable from reality. Recently, a video went viral showing a supposed Ukrainian soldier crying on camera, which was actually generated by AI using the face of a Russian content creator critical of Putin. These images, like others of fake soldiers, show hardly any perceptible errors, except for small details in uniforms or insignias, making them difficult to detect. This contrasts with the more crude manipulations of the early months of the war, such as the deepfake of Zelensky dancing, which displayed obvious flaws. “Current generative technologies have reached such realism that identifying manipulation requires advanced analytical tools, as these attacks seek to influence the public's subconscious,” warns Petriv. The expert also warns about the risk that the proliferation of artificial images could serve as an excuse to deny real facts, as it is now possible to discredit authentic evidence simply by labeling it as “AI-generated.” This creates a dangerous presumption of falsehood, where truthful information is left unprotected against systematic denial. Some of these AI-generated images have been disseminated in several languages, including German, Spanish, French, English, Italian, and Portuguese, suggesting the existence of coordinated campaigns. Additionally, outside the realm of AI, in recent months, misinformation campaigns have been detected aimed at dissuading foreign fighters, especially Colombians, from joining Ukrainian ranks, and even obstructing peace negotiations.