Following the online dissemination of images depicting the aftermath of Hurricane Helene, a particular image showcasing a crying child holding a puppy on a boat attracted widespread attention. Several posts on X, previously known as Twitter, featuring this image garnered millions of views. The image sparked emotional reactions among many users, including numerous Republicans who criticized the Biden administration’s disaster response. However, others quickly identified the image as likely being created using generative artificial intelligence (AI) tools due to recognizable features such as malformed limbs and blurriness.
This trend of AI-generated synthetic images has increased on social media, particularly following politically contentious news events. Observers of online platforms and elections note that these images frequently serve to propagate partisan narratives, with factual accuracy often taking a backseat. After users on X labeled the image of the child on the boat as likely being AI-generated, some individuals who had shared the image, such as Senator Mike Lee from Utah, removed their posts. Meanwhile, others continued to share the image. Republican National Committee member Amy Kremer, representing Georgia, stated that the source of the image was not of concern.
Renée DiResta, a professor at Georgetown University’s McCourt School of Public Policy, described the phenomenon as a form of political propaganda and a method for declaring interest and support for candidates. According to her, political campaigns often amplify such narratives by retweeting and boosting them.
In addition to the child image, other AI-generated images depicting animals on flooded roofs emerged after hurricanes Helene and Milton. Some AI-generated images spread on social media, such as those featuring former President Trump cuddling cats and ducks, following unfounded claims regarding Haitian immigrants in Ohio. DiResta points out that generative AI offers a straightforward and engaging way for campaign supporters to interact with political narratives online.
The debate surrounding truth versus facts in images is prominent in discussions about AI-generated content. For example, Amy Kremer defended the use of the synthetic image, claiming it symbolized the trauma and distress affecting many individuals. Matthew Barnidge, a University of Alabama professor, highlights how Western philosophical concepts such as those of Kant and Kierkegaard revolve around the idea of a deeper truth, distinct from mere facts.
Research suggests that while fact-checking can impact voters’ perceptions of politicians, verifying images presents more challenges than verifying text. Emily Vraga, a health communication researcher at the University of Minnesota, emphasizes that identifying the authenticity of online images has always been difficult, even before the advent of advanced AI like ChatGPT.
Hyper-realistic AI-generated images straddle the line between fact and fiction, often leaving a lasting impression. New York University art historian Ara Merjian notes that a photorealistic AI-generated image of pop star Taylor Swift endorsing Trump resulted in public reactions, highlighting the visual medium’s powerful influence. Swift later cited misinformation, including AI-generated content, as a reason for endorsing Vice President Kamala Harris.
AI-generated images have not only influenced political narratives but also filled voids left by the traditional news industry. Barnidge suggests that this space is being occupied by propagandists who blend into lifestyle content rather than presenting as traditional news. Researchers have observed AI-generated content depicting varied themes, from cakes and animals to dramatic religious imagery.
An investigation by 404 Media revealed how individuals in developing countries utilize AI-generated images to create trending posts on platforms like Facebook, thereby earning financial rewards. These content farms produce images that evoke emotions, often patriotic or artistic in nature.
One notable politically-related AI image, shared by X’s owner Elon Musk, depicted a figure resembling Vice President Kamala Harris in a red uniform with a hammer and sickle insignia. Eddie Perez, of the OSET Institute, suggested that images like this aim to depict Harris as un-American and undermine Democratic electoral legitimacy. According to Perez, such images can exacerbate political polarization and potentially deceive the public regarding electoral integrity. He noted that while generative AI represents one of many tools facilitating these strategies, it does not singularly elevate or escalate their impact but remains a contributing factor.
