The Claim

On June 5, 2026, social media accounts on X, Facebook, TikTok, and Threads circulated an image showing a blue-haired person dressed in all black, holding multiple folded papers in front of a ballot drop box. Posts framed the image as proof of "ballot harvesting" in California, with some claiming it was photographed in Glendale. The claim was widely amplified during the contested Los Angeles mayoral election. PolitiFact investigated the image on June 8, 2026 and determined it is Synthetically Generated — produced using Google's Gemini AI tool, not captured at any California ballot drop location.

Key Evidence

  • Gemini watermark confirmed: The image contains the visible logo of Google's Gemini AI tool in the bottom-right corner. Google's AI-generated images embed SynthID, a digital watermark invisible to the human eye but detectable by Google's verification technology. When PolitiFact uploaded the image to Gemini, the system confirmed the image was "edited or generated with Google AI." This is the most direct and technically definitive finding in the investigation.
  • Warped text and misspellings on the drop box: Frame analysis of the image revealed that lettering on the ballot drop box in the background contains warped characters and misspellings — a hallmark failure of AI image generation, which struggles to render coherent text. This artifact is consistent with generative AI output and inconsistent with photographs of real physical objects.
  • Zero corroborating incident reports: PolitiFact found no news reporting from any California outlet, no law enforcement reports, and no official statements describing a ballot stuffing incident of this kind at any drop box location during the current election cycle. Genuine incidents of this nature — if they occurred — would generate immediate documentation given the saturation of surveillance cameras around California drop boxes.
  • Broader misinformation context: The image spread in direct conjunction with President Trump's June 7, 2026 interview on NBC's Meet the Press, during which he falsely claimed that ongoing California ballot counting meant "they're cheating on the election." PolitiFact separately rated that claim "Pants on Fire." The AI image served as apparent visual evidence for a narrative that pre-existed any specific incident.

California Drop Box Security Procedures

California law mandates specific security measures for all official ballot drop boxes. They are required to be placed in publicly accessible locations and monitored by continuous video surveillance. The state permits voters to designate another person to return their ballot — a practice commonly called "ballot collection" that is legal under California election code and subject to strict chain-of-custody documentation requirements. Drop boxes are sealed and retrieved by election officials, not bystanders. This context matters because much of the viral framing of the image implied that what it depicted would constitute illegal activity; even setting aside the fact that the image is fabricated, the scenario it depicts does not straightforwardly represent the crime it purports to show.

Election Misinformation Pattern

AI-generated visual content has become an increasingly common instrument for injecting false evidence into election narratives. The California mayoral race provided a high-attention context in which fabricated imagery could gain rapid traction — particularly among audiences already primed by political messaging to expect irregularities. The technical markers left by Gemini — the watermark, the SynthID signature, the text distortions — are traceable, but most users sharing the image lacked the tools or incentive to check them. This investigation demonstrates how swiftly synthetic imagery can function as apparent proof of a claim that has no basis in documented reality.

Veredicto also investigated this fabrication. Their full analysis is available at Veredicto.

Verdict

The image circulating on social media does not show a real person stuffing ballots at a California drop box. It is Synthetically Generated — created using Google's Gemini AI tool, as confirmed by a visible Gemini watermark, SynthID detection, distorted text artifacts, and the complete absence of any corroborating real-world incident. PolitiFact first reported these findings on June 8, 2026. The image is fabricated evidence deployed in an existing election fraud narrative.