Documents offer rare insight on Ice's close relationship with Palantir (www.theguardian.com)

A cache of internal ICE documents obtained via FOIA shows Palantir’s analytics platforms have been deeply embedded in US Immigration and Customs Enforcement workflows since 2014, contradicting the company’s frequent framing as a passive “data processor.” The records—emails, training manuals and reports covering 2014–2022—detail use of Palantir-built tools like Falcon (a desktop/mobile app) and Investigative Case Management (ICM) to ingest and search across vast federal and commercial datasets (EID, SEVIS, APIS, CBP border data, subscription services such as Thomson Reuters Clear), phone extracts unlocked with Cellebrite, driver’s-license scans, cell-tower and GPS “blueforce” tracking, and real-time field reporting. Palantir engineers also embedded with ICE teams to create new use cases and operationalize features for raids, airport and event deployments; Falcon was used for planning, location-tracking and on-the-ground intelligence until HSI replaced it with an internal tool in 2022. Palantir recently won a $30M contract to build “ImmigrationOS,” expanding these capabilities to ICE’s enforcement arm (ERO). For the AI/ML community, the documents are significant because they reveal how scalable data-integration platforms and real-time analytics are being applied to mass surveillance and deportation operations. Technically, Palantir’s systems demonstrate powerful cross-database linking, mobile-enabled evidence ingestion and rapid entity resolution at scale—capabilities that raise urgent questions about data governance, model bias, provenance, auditing, and lawful-purpose constraints. As ICE gains access to additional feeds (IRS records, Clearview-style facial recognition) and tools for automation, researchers and practitioners must weigh the societal risks of building and deploying analytics infrastructure that can be repurposed for wide-reaching state surveillance and enforcement.