ICE is threatening to deport witnesses of its latest shooting

Even with three eyewitness accounts that contradict its story, the agency is using a self-defense justification for the fatal shooting of Lorenzo Salgado Araujo.


Advocates are demanding that the Department of Homeland Security release bodycam footage of the fatal shooting of Lorenzo Salgado Araujo, a Mexican immigrant who was killed by ICE officers in Houston during a traffic stop earlier this week. But DHS claims the agents involved in the shooting weren’t wearing body cameras because of the lengthy government shutdown that prevented ICE and Customs and Border Protection from receiving additional federal funding for 76 days — a shutdown that was itself spurred by congressional bickering over reforms to DHS after federal agents killed two civilians earlier this year.
Two competing narratives have emerged in the absence of any footage of the shooting, which occurred Tuesday around 7AM as Salgado Araujo, who owned a construction business, drove to a work site with three of his employees. But as the killings of Renee Good and Alex Pretti demonstrated, DHS will readily accuse victims of attacking agents — even when video evidence shows otherwise.
But what good is video footage or eyewitness accounts in a world where the government will just deny accountability, even when confronted with evidence that clearly contradicts their statements? Even if footage of Salgado Araujo’s shooting does emerge, it won’t stop DHS from saying whatever it wants about the people it kills.
An ICE spokesperson said Salgado Araujo “weaponized his vehicle in an attempt to run over an ICE law enforcement officer” — which echoes claims DHS has made after other shootings, some of them fatal. But three eyewitnesses, all of whom are in ICE detention, maintain that Salgado Araujo did no such thing. According to their accounts, which attorney Hugo Balderas-Ibarra shared with The Washington Post , ICE vehicles surrounded the work van on both sides and shot into their vehicle. Juan Proaño, the CEO of the League of United Latin American Citizens, told The New Republic that DHS is pressuring the witnesses to self-deport.
In a statement released after the shooting, DHS said agents stopped Salgado Araujo “as part of a targeted enforcement operation.” But sources with knowledge of the situation told both The New York Times and CNN that Salgado Araujo wasn’t the target. Officers were reportedly looking for two Guatemalan men. While surveilling a property connected to the Guatemalan men, officials had seen two white vans, a DHS spokesperson told the Times . Later, “they observed a white van with an individual who resembled the target” — the van driven by Salgado Araujo.
Verified source · The Verge
Reported by The Verge. Open the original for full media and formatting.
More in Agents
All newsRemember When It Matters: Proactive Memory Agent for Long-Horizon Agents
arXiv:2607.08716v1 Announce Type: new Abstract: In long-horizon tasks, decision-relevant state is often scattered across an expanding trajectory, while the action agent must surface it and act. As trajectories grow, task requirements, environment facts, prior attempts, diagnoses…
Read at arXiv cs.AIFormal Mechanisms for Market Stability in Self-Interested Agent Societies: A Marketplace Simulation Study
arXiv:2607.08652v1 Announce Type: new Abstract: Self-interested agents, left unconstrained, tend toward defection in repeated social dilemmas, causing cooperative gains from trade to collapse. This paper investigates what formal mechanisms, layered on top of unrestricted communi…
Read at arXiv cs.AIFrom Prompts to Contracts: Harness Engineering for Auditable Enterprise LLM Agents
arXiv:2607.08028v1 Announce Type: new Abstract: Enterprise large language model (LLM) applications often begin as prototypes whose behavior is carried by prompts and retrieval context. Productization adds requirements for source boundaries, entity routing, answer contracts, and…
Read at arXiv cs.AIAUTOPILOT VQA: Benchmarking Vision-Language Models for Incident-Centric Dashcam Understanding
arXiv:2607.08745v1 Announce Type: new Abstract: Recent advances in Vision-Language Models, Large Language Models, and Multimodal Large Language Models have improved autonomous driving tasks such as scene understanding, decision making, trajectory prediction, and visual question…
Read at arXiv cs.AI