The US Immigration and Customs Enforcement (ICE) has used an AI-powered data-scanning tool called Giant Oak Search Technology (GOST) to scour social media looking for post containing "derogatory" comments about the nation.

The GOST database ranks social media scores from one to 100, and is searchable by name, address, email address, and country of citizenship. Immigration agencies and service providers have apparently been using the data in enforcement actions, according to an The American Civil Liberties Union lawsuit filed under the Freedom of Information Act (FOIA), and first reported by 404 Media.

ICE is one of seven federal agencies that surveils social media to monitor users' speech online. The ACLU delivered those agencies an FOIA request in 2018. None of the seven handed over any documents, and in January 2019, the ACLU filed a lawsuit to force them to turn over the requested information.

"The Biden administration has been quietly deploying and expanding programs that surveil what people say on social media, often without any suspicion whatsoever," Shaiba Rather, a Nadine Strossen Fellow with ACLU's National Security Project, told The Register. "These programs chill people from speaking freely online and transform social media into a platform for constant government scrutiny."

"Our ongoing FOIA lawsuit exposes the black box systems like GOST that have enabled this surveillance," Rather continued. "The government shouldn't be conducting suspicionless monitoring of our online speech - and it certainly shouldn't be doing so in secret."

Giant Oak, the company that developed GOST, bills itself as seeing "the people behind the data," according to its website. The firm says its AI-based system allows government agencies and law enforcement to "identify bad actors by behavioral pattern rather than identity labels," using information found on the open and deep web.

GOST customers, according to the company, include several US government bodies: Department of Defense, Department of Homeland Security, Department of State, Department of Treasury, and law-enforcement agencies.

"GOST gives government agencies charged with the protection of US borders the ability to rapidly screen for terrorist affiliations or violent criminal histories the hundreds of thousands of foreign nationals entering the US border each day," the website states.

DHS has reportedly used GOST since 2014, according to documents obtained by 404 Media, and ICE has paid Giant Oak more than $10 million for the system since 2017.

Other documents detail analysts using GOST to review "potentially derogatory social media," giving images either a "thumbs up" or "thumbs down" rating, and reportedly scanning the individual's social media profile to understand their connections to other social media users.

Giant Oak was also involved in a 2016 pilot called the "HSI [Homeland Security Investigations] PATRIOT Social Media Pilot Program," according to the FOIA documents, which intended to monitor social-media posts to assist agents in making immigration-related decisions.

"The premise of this initiative is to track non-immigrants from the time they file a visa application with the Department of State, to the time they enter the United States, and through the time when they either depart the United States, or until such time as they become an overstay or otherwise fail to comply with their terms of admission," according to a DHS document [PDF].

The Giant Oak contract with DHS ended in August 2022, according to the records.

ICE and Giant Oak did not respond to The Register's request for comment. ®

https://www.theregister.com//2023/10/25/ice_social_media_surveillance/