AI-generated deepfake images of Donald Trump being arrested were created and spread widely on social media, causing confusion and demonstrating the potential for synthetic media to mislead the public during significant news events.
In March 2023, as former President Donald Trump faced potential indictment over hush money payments to adult film star Stormy Daniels, AI-generated deepfake images depicting his arrest flooded social media platforms. Eliot Higgins, founder of investigative collective Bellingcat, created many of the viral images using Midjourney, an AI text-to-image generator, with prompts like 'Donald Trump falling over while getting arrested. Fibonacci Spiral. News footage.' The fabricated images showed Trump resisting arrest by NYPD officers, running from police, and being restrained, with some depicting family members protesting his arrest. While Higgins clearly labeled the images as AI-generated and created them as commentary, they were viewed over 5 million times and were subsequently shared by others without context, with some users falsely presenting them as genuine photographs. The images contained telltale signs of AI generation including distorted hands, garbled text on police uniforms, and anatomical inconsistencies like Trump appearing with three legs or wearing a police belt. The NYPD confirmed no such arrest had occurred. The incident highlighted concerns about the increasing sophistication of AI-generated imagery and its potential to spread misinformation during crisis events, with experts warning about the erosion of trust in visual media.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed