Leaked records from the Department of Homeland Security reveal extensive funding for artificial intelligence surveillance projects. The Office of Industry Partnership distributed contracts to over 1,400 companies aimed at expanding monitoring capabilities. This disclosure follows a significant $165bn funding increase allocated to the agency in the previous year, raising concerns about oversight. The data suggests a rapid acceleration in automated tracking technologies across federal jurisdictions, particularly regarding civil liberties.
The leaked database includes over 6,000 companies that bid for contracts without necessarily receiving funding. Specific projects cover automated airport surveillance and mobile biometric scanning tools for agents. One initiative proposes an AI platform to ingest 911 call data and generate predictive crime maps. These tools represent a shift toward centralized data processing for public safety incidents and national security strategies.
Data obtained by a pseudonymous cyber-hacktivist was provided to reporters by Distributed Denial of Secrets. The material spans two decades of contracting history from 2004 through late 2025. Records indicate contracts worth a combined $845m were awarded for prototype development and research. The Guardian verified portions of the data against existing public records before publication to ensure accuracy.
Recent contracts issued in May 2025 focus on enabling agents to harvest biometric data using cellphones. Companies like Intellisense Systems and Integrated Biometrics received funding to create handheld scanning adapters. These devices promise compatibility with standard Android and iOS operating systems without custom drivers. The technology aims to extend DHS capabilities beyond traditional fixed locations and into field operations.
Airport security proposals involve AI systems analyzing CCTV feeds to catalog passenger physical characteristics. Four contracts totaling $699,000 were awarded to firms including Synthetik and Toyon Research Corporation. These systems claim to detect clothing types and accessories to flag potential threats automatically. Critics note the TSA previously spent $900m on a behavioral screening program that failed to deliver results.
A separate initiative by Cassius LLC proposes ingesting data from 5,000 national 911 call centers. The platform intends to build geospatial heat maps to predict incident trends using AI models. This approach mirrors predictive policing methods previously abandoned by cities like Los Angeles and Chicago. The Brennan Center for Justice describes such technology as tech-washing for biased policing methods.
Jeramie Scott of the Electronic Privacy Information Center criticized the strategy as following dystopian fiction tropes. He stated the department is taking the wrong lesson from science fiction narratives regarding surveillance. Previous attempts at technology-assisted behavioral screening were wound down following disclosures of racial profiling. Officials argue new AI tools offer more objective analysis than human observation alone, despite historical failures.
Responses from the involved companies vary from silence to unprofessional dismissals of inquiry attempts. Zachary Canders of Cassius LLC responded to The Guardian with an informal email declining substantive comment. DHS officials have not yet provided a detailed justification for the expanded surveillance scope. Privacy advocates warn these tools could undermine democratic processes if used improperly by law enforcement.
