Concerns Over AI-Generated Imagery in Political Context

The Trump administration has actively embraced AI-generated imagery, sharing cartoonlike visuals and memes through official channels. However, a realistic altered image of civil rights attorney Nekima Levy Armstrong in tears, following her arrest, has raised concerns about the blurred lines between real and fake imagery.

Homeland Security Secretary Kristi Noem’s account originally posted the image of Levy Armstrong’s arrest before the White House account shared a doctored version showing her crying. This image is part of a widespread use of AI-edited imagery across the political spectrum, particularly following the fatal shootings of Renee Good and Alex Pretti by U.S. Border Patrol officers in Minneapolis.

Misinformation experts are troubled by this development, fearing that the spread of AI-generated or edited images damages public perception and breeds distrust. In response to criticism, the White House doubled down, with Deputy Communications Director Kaelan Dorr writing on X, “memes will continue.” Deputy Press Secretary Abigail Jackson also mocked the criticism.

David Rand, a Cornell University professor of information science, noted that calling the altered image a meme might be an attempt to present it as a joke, similar to previous cartoonish posts. He expressed ambiguity about the purpose behind sharing the edited image, contrasting it with past humorous content from the administration.

Zach Henry, a Republican communications consultant, stated that AI-enhanced or edited imagery is a tool used by the White House to engage tech-savvy segments of Trump’s online base. He suggested that such imagery can go viral if it provokes strong reactions.

Michael A. Spikes, a Northwestern University professor, emphasized the importance of the government providing accurate information. He warned that sharing manipulated content erodes public trust in the federal government, exacerbating existing distrust in news organizations and higher education.

Ramesh Srinivasan of UCLA highlighted the growing challenge of finding “trustable information.” He pointed out that AI systems amplify issues of trust and reality understanding. Srinivasan warned that the White House’s use of AI-generated content might embolden others in positions of power to share synthetic content.

An influx of AI-generated videos related to Immigration and Customs Enforcement (ICE) actions and protests has proliferated on social media. The shooting of Renee Good by an ICE officer led to numerous AI-generated videos depicting dramatic encounters with ICE agents.

“It’s going to be an issue forever now,” said Jeremy Carrasco, a media literacy expert. He doubts people understand the severity of the situation.

Carrasco suggests a watermarking system that embeds information on media origin into metadata as a potential solution. However, widespread adoption is still far off.

Fabricated images following events like the capture of Nicolás Maduro have also surged online, signaling a growing trend for AI-generated political content.

Associated Press writers Jonathan J. Cooper and Barbara Ortutay contributed to this report.