DHS Confirms Use of Google, Adobe AI in Public Content Creation

The US Department of Homeland Security (DHS) is harnessing cutting-edge generative AI tools from Google and Adobe to craft and disseminate videos and images for public communication, according to a recently published inventory of government AI use. This disclosure provides the first concrete evidence that federal agencies are actively deploying AI-generated multimedia in their outreach, a revelation that is both technically significant and politically sensitive.

Expanding Role of AI in Government Messaging

The document, released by DHS on Wednesday, details the department’s adoption of commercial AI products for a variety of internal and public-facing tasks. Notably, it confirms the use of Google’s Veo 3 video generator and Adobe’s Firefly platform for creating and editing visual materials, with DHS reportedly holding between 100 and 1,000 licenses for these advanced systems.

The AI tools in question are among the most sophisticated available to date. Google’s Veo 3, for instance, allows users to generate hyperrealistic video clips complete with sound, dialogue, and background ambiance. Adobe’s Firefly, first launched in 2023, is marketed as an AI generator that avoids the use of copyrighted material both in its training data and its outputs. Both platforms offer functionalities that go beyond simple video editing, enabling the automated assembly of entire multimedia packages—including images, soundtracks, and speech—intended for public dissemination.

Reinforcing Political Agendas

The news comes amid a surge of AI-generated content circulating on US government social media channels, particularly those promoting President Trump’s immigration and deportation policies. Agencies within DHS, such as Immigration and Customs Enforcement (ICE), have released videos and posts referring to religious holidays, showing the faces of detainees, and running recruitment advertisements—all at a volume and pace that suggest technological augmentation.

For months, observers and rights advocates have speculated about the extent of AI use in such content, noting telltale signs of synthetic imagery and audio. However, ambiguity persisted over which, if any, specific AI systems were behind these productions. The newly disclosed inventory bridges this gap, confirming for the first time that both Google’s and Adobe’s AI products play a role in the agency's public-facing communication.

Challenges in Identifying AI-Generated Content

Despite these revelations, the provenance of much online government material remains murky. While Adobe does permit watermarks to signal AI-generated media, such identifiers are not always retained as content travels across platforms, making consistent transparency a challenge.

The document further reveals that DHS utilizes Google’s Flow, an integrated toolset combining Veo 3 with additional filmmaking capabilities, underscoring the growing sophistication of in-house content creation by government agencies. In addition to these generative platforms, the department has adopted Microsoft Copilot Chat for drafting documents and working with reports, as well as Poolside’s coding tools for software tasks.

Ethical Tensions Within Tech Companies

The DHS partnership with major tech firms has drawn criticism from within the technology sector itself. In recent weeks, more than 170 employees from Google and Adobe have publicly urged their employers to reconsider their relationships with US immigration authorities, particularly in light of reported uses of AI-multimedia in operations with humanitarian implications. Corporate leaders at Google and Adobe have so far refrained from commenting on the controversy.

Meanwhile, both companies have made moves to distance themselves from direct facilitation of controversial law enforcement actions, such as the October removal (by Google and Apple) of mobile apps designed to track ICE activity, citing safety and privacy concerns.

Broader Implications and Ongoing Questions

Another document released alongside the AI-tool inventory describes additional technological deployments, including a new facial recognition application reportedly in use at ICE. This further highlights the increasingly central role of AI-driven tools in government operations ranging from recruitment to enforcement and identification.

The expansion of generative AI in government, particularly in areas as contentious as immigration policy, raises pressing questions regarding transparency, the potential for manipulation, and safeguards around the use of synthetic media. While the European Union is advancing comprehensive regulatory frameworks—such as the AI Act—to address such issues, the US government’s strategy appears focused on transparency through disclosure, even as it harnesses these powerful tools.

For European observers, the US experience underscores the necessity of robust oversight and public debate as AI adoption deepens within the public sector, setting an international context for ongoing discussions about transparency, ethics, and accountability in government use of artificial intelligence.

Reference: DHS is using Google and Adobe AI to make videos