The Trump administration has openly shared AI-generated visuals on official channels, such as cartoon-like images and memes. Recently, a realistic edited image of civil rights attorney Nekima Levy Armstrong cried after her arrest stirred further concerns about blurring the lines between fiction and reality.
Homeland Security Secretary Kristi Noem initially posted an image from Levy Armstrong’s arrest, but later the White House shared an altered version showing her in tears. The use of AI-edited images has become increasingly prevalent since the fatal shootings of Renee Good and Alex Pretti by U.S. Border Patrol officers in Minneapolis. Experts worry this practice erodes public trust and spreads misinformation.
In response to criticism, White House officials defended the post, with Deputy Communications Director Kaelan Dorr stating that memes will persist. Deputy Press Secretary Abigail Jackson also mocked the critique, fueling the controversy.
David Rand, an information science professor at Cornell University, points out that labeling the altered image as a meme may be an attempt to cast it as humorous, similar to previous posts. This tactic might be intended to deflect criticism.
Although memes convey layered messages, AI-enhanced imagery is a new tool used to engage Trump’s base, particularly active online users. Republican communications consultant Zach Henry mentions that while digitally-savvy people recognize it as a meme, it may confuse others, potentially sparking reactions that aid viral spread.
Michael A. Spikes, a Northwestern University professor, emphasizes the impact of sharing distorted images by credible sources. Rather than accurately representing events, fabricated content crystallizes false perspectives. He argues the government should provide trustworthy, verified information, highlighting a significant trust issue, particularly amid institutional crises regarding distrust in news organizations and higher education.
Ramesh Srinivasan from UCLA raises concerns over the long-term effects of AI systems worsening the absence of trust in genuine information. When officials share AI-generated content, it encourages similar posts from everyday users and from figures with authority, like policymakers. Given social media’s tendency to favor extreme or conspiratorial content, AI generation tools introduce complex challenges.
AI-edited imagery and videos have already proliferated online, particularly regarding Immigration and Customs Enforcement actions. Following Renee Good’s shooting by an ICE officer, several fabricated videos emerged featuring women fleeing ICE officers.
Jeremy Carrasco, a media literacy expert, warns that many viewers can’t differentiate between real and fake content. While obvious signs of AI use exist in the best scenarios, viewers may not notice.
This issue extends beyond immigration enforcement news, with fake images about Venezuelan leader Nicolás Maduro’s capture circulating recently. Experts foresee increased AI-generated political content.
To address these challenges, a watermarking system embedding media origin metadata has been developed by the Coalition for Content Provenance and Authenticity. Carrasco suggests widespread adoption may take time.
Associated Press writers Jonathan J. Cooper in Phoenix and Barbara Ortutay in San Francisco contributed to this report.