A newly released US Department of Homeland Security document has provided a clearer view of the AI systems the agency uses to create and edit public-facing material. The inventory says DHS uses Google’s Veo 3 video generator and Adobe Firefly for work involving images, videos, and public affairs content.
The disclosure matters because immigration agencies have been posting a large volume of social media content as immigration operations have expanded across US cities. Some of that content appears to be AI-generated, although the document does not identify which tool was used for any specific post.
What the DHS document reveals
The document, released on Wednesday, lists commercial AI tools used across DHS for a range of tasks. Those tasks include drafting documents, summarizing long reports, supporting coding work, managing cybersecurity, and editing public affairs materials.
In the section focused on “editing images, videos or other public affairs materials using AI,” the document names Google’s Veo 3 and Adobe Firefly. It estimates DHS has between 100 and 1,000 licenses for the tools.
The inventory also says DHS uses Microsoft Copilot Chat to generate first drafts of documents and summarize long reports. It lists Poolside software for coding tasks, along with tools from other companies.
Google, Adobe, and DHS did not immediately respond to requests for comment, according to the source article.
Why the video tools are drawing attention
The new information gives more detail about how agencies within DHS, including Immigrations and Customs Enforcement, might be producing material for X and other public channels. The document is significant because it is the first concrete evidence that DHS is using these AI video generators to create content shared with the public.
ICE and related agencies have posted content tied to President Trump’s mass deportation agenda. The posts have included material celebrating “Christmas after mass deportations,” references to Bible verses and Christ’s birth, faces of people the agency has arrested, and ads aimed at recruiting agents.
The agencies have also repeatedly used music in videos without permissions from artists. Separately, some of the videos have the appearance of being AI-generated. Until this document, however, it had not been clear which AI models DHS might be using for public-facing video work.
The document does not prove that any one video was created with Google’s or Adobe’s systems. It also does not show exactly how DHS staff are using those products or which agency office is responsible for particular posts.
What Google Flow and Adobe Firefly can do
The DHS document specifically says the agency is using Flow, a Google tool that combines Veo 3 with filmmaking features. Flow can be used to generate clips and assemble full videos with AI. The source article notes that these videos can include sound, dialogue, and background noise, which can make them hyperrealistic.
Adobe launched Firefly in 2023. The company has said Firefly does not use copyrighted content in its training or output. Like Google’s tools, Firefly can generate videos, images, soundtracks, and speech.
Those capabilities help explain why the DHS disclosure is being scrutinized. Tools that can create complete video scenes, audio, and speech can make it harder for the public to understand how a government message was produced. At the same time, the document does not disclose operational details about how DHS applies the tools in practice.
Verification remains difficult
Even with the inventory, it remains impossible to verify which company’s system helped create a specific piece of content. It may also be impossible to confirm whether a particular video was AI-generated at all.
Adobe offers options to “watermark” a video made with its tools to disclose that it is AI-generated. But the source article notes that such disclosure does not always remain intact after content is uploaded and shared across different sites.
That creates a gap between internal tool use and public visibility. DHS may have access to AI video generators, and public content may look synthetic, but viewers cannot reliably connect any single post to any single model based on appearance alone.
Pressure inside tech companies
The disclosure also arrives as workers at large technology companies are pressing their employers over immigration enforcement. In recent weeks, more than 140 current and former employees from Google and more than 30 from Adobe have pushed their companies to take a stance against ICE and the shooting of Alex Pretti on January 24.
Google’s leadership has not made statements in response. In October, Google and Apple removed apps from their app stores that were intended to track sightings of ICE, citing safety risks.
A separate document released on Wednesday also provided new details about DHS use of more niche AI products. That includes a facial recognition app used by ICE, as first reported by 404Media in June.
Taken together, the documents show that DHS use of AI is not limited to one public communications tool. The agency inventory includes systems used for public media, document work, coding, cybersecurity, and other specialized tasks. For the public, the most immediate question is how clearly AI-generated government communications will be labeled when they appear on social platforms.