The Trump administration has increasingly shared AI-generated imagery online, using cartoonlike visuals and memes on official White House channels. A recent incident involving an altered image of civil rights attorney Nekima Levy Armstrong has sparked controversy. The edited image, showing her in tears after an arrest, has alarmed many who worry about the blurred lines between reality and manipulation.
Homeland Security Secretary Kristi Noem initially posted the original image of Levy Armstrong’s arrest. Soon after, the White House shared the doctored version, displaying her crying. This incident is just one among many AI-edited images circulated since the fatal shootings of Renee Good and Alex Pretti by U.S. Border Patrol officers in Minneapolis. Experts express concern that the use of AI to alter images could erode trust and perpetuate misinformation.
White House officials maintained their position despite criticism, with Deputy Communications Director Kaelan Dorr stating on X that memes “will continue.” Deputy Press Secretary Abigail Jackson shared a post mocking the critique.
David Rand, an information science professor at Cornell University, views the altered image as an attempt to present it as humorous. According to Rand, the purpose of sharing such an image is less clear than previous cartoonish memes. These types of media carry layered messages that are familiar to certain audiences but confusing to others.
Zach Henry, a Republican communications consultant, explains that memes are recognized by online communities while appearing real to those less familiar with the context, like older generations. The reaction, whether negative or positive, can help make the content viral.
Michael A. Spikes, a professor at Northwestern University, argues that altered images from credible sources can distort public perception. The government should provide trustworthy information. Sharing manipulated content undermines this responsibility, adding to the existing distrust in institutions.
Ramesh Srinivasan, a UCLA professor, highlights the broader implications of AI-generated content. He warns that it amplifies challenges related to trust and blurred realities. Srinivasan believes that when officials use AI-generated content, it encourages others in power to do so, presenting dangers like privileging extreme content on social media.
A surge of AI-generated videos related to ICE actions illustrates this problem. After an ICE officer shot Renee Good, videos depicting confrontations with ICE officers appeared online, adding to the confusion. Content creator Jeremy Carrasco points out that these videos often aim for engagement through sensational keywords.
Despite evident signs of AI generation, such as random street signs, viewers struggle to discern authenticity. Carrasco argues that in situations of higher stakes, determining what’s real remains critical. He suggests implementing watermarking systems that track media origins, although widespread adoption is uncertain.
Experts, including Carrasco, anticipate a growing prevalence of AI-generated political content. Measures to address this issue, such as watermarking, are still in development. Meanwhile, the use of AI-enhanced content poses ongoing challenges to public trust and media literacy.