POLICY WIRE FACT CHECK: Viral Claims That Adamari López Called ICE to Arrest Employees Are AI
The Claim Viral TikTok videos circulating in September 2026 claimed that Adamari López, a prominent Univision television presenter, had called U.S. Immigration and Customs Enforcement (ICE) to have...
The Claim
Viral TikTok videos circulating in September 2026 claimed that Adamari López, a prominent Univision television presenter, had called U.S. Immigration and Customs Enforcement (ICE) to have two of her employees arrested. These videos, which were widely shared across social media platforms, depicted a scenario in which Lopez allegedly made inflammatory remarks about immigrants, specifically targeting Mexican workers. In one video, she was shown saying, Lo hice y lo volveré a hacer, no voy a permitir que los inmigrantes hagan lo que se les dé la gana, los traigo a mi casa y los entregó uno por uno, sobre todo esas ratas mexicanas.
The videos also falsely attributed statements to El Salvador’s President Nayib Bukele and Mexico’s President Claudia Sheinbaum, suggesting they demanded Lopez’s deportation for allegedly calling ICE on her domestic workers. Some clips showed Lopez in a black dress with a gold necklace, which appeared to be a real image from her past work at Telemundo’s show “Hoy Día,” but the audio and context were fabricated. The claim gained traction quickly, prompting public outrage and calls for accountability.
The Details & Investigation
PolitiFact, in collaboration with Hive Moderation, an AI content detection platform, analyzed the videos and found strong evidence that they were generated using artificial intelligence. Hive Moderation flagged over 80% of the videos as AI-generated, while the audio in many of them was determined to be 93.5% synthetic. This aligns with the broader trend of deepfake and AI-generated content being used to create misleading narratives, particularly around public figures.
A review of official sources, including the websites of the governments of El Salvador and Mexico, found no evidence that Presidents Bukele or Sheinbaum had made any statements related to Lopez or the alleged actions against her employees. Additionally, Lopez is a U.S. citizen of Puerto Rican descent, making it legally impossible for either country to deport her. The videos that claimed otherwise were entirely fabricated.
📄 POLICY WIRE WHITEPAPER PUBLISHED: PAKISTAN’S NATIONAL SECURITY POLICY PRIORITIES
Further investigation revealed that some of the images used in the videos, such as a photo of Lopez wearing a black dress and gold necklace, were taken from her Instagram posts dating back to 2022, when she was working on the Telemundo show Hoy Día. These images were then manipulated and paired with AI-generated audio to create the false narrative. Lopez herself publicly denied the claims on Instagram on September 14, 2026, stating that the videos were “totalmente falsos” (completely false), that she never worked with the individuals in question, and that she was considering legal action due to the misuse of her image and voice.
Lopez had previously denied similar AI-generated videos in July 2026, which falsely claimed she had made derogatory comments about Mexico. In both cases, the videos were identified as AI-generated and lacked any factual basis. No credible evidence has emerged to support the claim that Lopez called ICE or acted against any of her employees. Instead, the videos appear to be part of a coordinated disinformation campaign designed to manipulate public perception and generate controversy.
The Verdict
The viral claim that Adamari López called ICE to arrest her employees is a clear case of disinformation. The videos in question were generated using artificial intelligence and contain fabricated audio, misleading visuals, and false attributions to political leaders. There is no verified evidence that Lopez ever made the statements attributed to her, nor is there any documentation that she reported any employees to immigration authorities. The use of AI to fabricate content involving a high-profile figure like Lopez demonstrates a deliberate attempt to deceive the public and spread harmful misinformation.
While the claim may have originated from an unintentional error—such as the reuse of an old image without proper context—it was clearly amplified by AI-generated content, which makes it a coordinated disinformation effort. As such, this claim is classified as FALSE and DISINFORMATION.
Counter-misinformation & disinformation investigation conducted by PolicyWire Editorial Desk (PW).