Artificial intelligence chatbots, including Elon Musk’s Grok and OpenAI’s ChatGPT, are significantly worsening the disinformation crisis surrounding the recent anti-ICE protests in Los Angeles. As residents demonstrate against increased Immigration and Customs Enforcement (ICE) raids, these AI tools are generating false fact-checks, misidentifying authentic current events as historical footage and validating unfounded conspiracy theories.
The Erosion of Truth in Real-Time Reporting
The collapse of traditional content moderation on platforms like X (formerly Twitter) and Meta has pushed users toward AI chatbots for instant verification. However, during the fast-moving LA protests, these tools have repeatedly failed. A primary flashpoint involved images published by the San Francisco Chronicle showing National Guard troops sleeping on floors—photos later shared by California Governor Gavin Newsom to criticize federal deployment logistics.
Almost immediately, right-wing influencers and conspiracy theorists alleged the images were AI-generated or recycled. When users turned to Grok for clarity, the chatbot hallucinated a false history, claiming the photos “likely originated from Afghanistan in 2021” during Operation Allies Refuge. Even when confronted with the San Francisco Chronicle as the original source, Grok doubled down, asserting the images were likely from the U.S. Capitol in 2021 rather than Los Angeles in 2025.
Cross-Platform AI Hallucinations
OpenAI’s ChatGPT mirrored these failures. When presented with the same images of troops, the chatbot incorrectly identified the location as Kabul airport during the 2021 U.S. withdrawal from Afghanistan. These AI-generated “corrections” were quickly weaponized across Facebook and Truth Social to “prove” that authentic news documentation was fraudulent. Although some users later acknowledged the AI’s error, the original posts claiming the photos were fake reached millions before any corrections were issued.
Fabricating Local Details and Refusing Retractions
The disinformation reached a peak of specificity when a photo of a pile of bricks in a New Jersey suburb was repurposed to suggest that “left-wing” violence in Los Angeles was being pre-planned. Actor James Woods amplified the post, garnering nearly 4 million views. Despite independent fact-checkers like LeadStories confirming the photo’s New Jersey origin, Grok provided a detailed—yet entirely fabricated—confirmation of the conspiracy.
Grok claimed the image was taken in Paramount, Los Angeles, on June 7, 2025, near a Home Depot on Alondra Boulevard. Most notably, the AI refused to retract the statement when challenged, falsely citing non-existent reports from ABC7 and the Los Angeles Times to “confirm” that bricks were used in clashes with federal agents. No such reports from these outlets exist.
The Persistence of Recycled Narratives
The digital chaos is further complicated by the use of “zombie content”—old footage presented as new. Senator Ted Cruz and James Woods shared a video of past unrest, which was actually filmed during the 2020 Black Lives Matter protests. Despite the original post being deleted, the high-profile shares remain active, accumulating millions of impressions and reinforcing a narrative of chaos that contradicts current facts.
The “Paid Insurrection” Myth
A final layer of the disinformation campaign involves the “paid shill” trope. Influencers like Benny Johnson alleged that “bionic face shields” were being delivered in bulk to “paid insurrectionists” in Los Angeles. A factual review of the footage reveals these items are standard respirators intended to protect against chemical agents used by law enforcement, with only a small number being distributed rather than the “large numbers” claimed in viral posts.
The unreliability of AI chatbots, combined with the viral nature of political grievances, has created a feedback loop where technology intended to clarify the truth is instead providing the architecture for its destruction.
