FOIA docs: ICE used a tool called GOST to decide whether US visa applicants' social media posts are “derogatory” to the US, as part of immigration enforcement
Check out this @404mediaco story on social media surveillance docs obtained by our @ACLU_NorCal team + @ACLU. https://www.404media.co/... Terri K O / @terri@social.afront.org : This article on how the US government uses black box algorithms on people's social media posts is a good time for a reminder that some of us, myself included, avoid talking about world events/politics in public not because we don't care but because we never know if it will be dangerous for us to voice any opinion. … Jason Koebler / @jasonkoebler@mastodon.social : Here's how ICE finds and analyzes “derogatory” posts about the US on social media. Then uses in immigration enforcement. — https://www.404media.co/... Taylor Lorenz / @taylorlorenz@mastodon.social : Joe Cox got a hold of internal documents that provide deep insight into GOST, a system used by ICE to find “derogatory” speech online. — Analysts can use a “thumbs up,” “thumbs down” decision on the material, and feed it into visa decisions and immigration enforcement https://www.404media.co/... @rgegriff@hackers.town : So some of the Vice Motherboard folks went independent after the layoffs; and they have been KILLING it lately with some of the investigative reporting. — https://www.404media.co/... X: Matt Cagle / @matt_cagle : The government should not be using algorithms to scrutinize our social media posts and decide which of us is “risky.” Check out this @404mediaco story on social media surveillance docs obtained by our @ACLU_NorCal team + @ACLU. https://www.404media.co/... Joseph Cox / @josephfcox : The tool is made by a company called Giant Oak. You may have never heard of them, but they've made millions providing social media monitoring to various US agencies. ICE, CBP, DEA, State, more. Here is an example of the UI. “Derogatory results” on the left https://www.404media.co/... [image] Joseph Cox / @josephfcox : We've seen lots of social media tools before, but what stands out is its explicit link to immigration, visas, etc. We have actual insight into the use case of this technology and the workflow https://www.404media.co/... [image] Emanuel Maiberg / @emanuelmaiberg : it feels like every other week joe uncovers some other wildly disturbing piece of software that ICE and DHS have used on anyone who comes close to the border https://www.404media.co/... Joseph Cox / @josephfcox : There appear to be a few different ways DHS is using the tool. One slide mentions the PATRIOT pilot, then another talks more specifically about overstay enforcement. Here are some of those slides. https://www.404media.co/... [image] Joseph Cox / @josephfcox : New: internal documents provide more insight into GOST, a system used by ICE to find “derogatory” speech online. Analysts can then use a “thumbs up,” “thumbs down” decision on the material, and feed it into visa decisions and immigration enforcement https://www.404media.co/... [image] Forums: BeauHD / Slashdot : ICE Uses Tool To Find ‘Derogatory’ Speech Online
Context & Ripple Effects
The FOIA records turn a broader pattern of immigration data collection into a specific decision workflow: ICE used GOST to assess visa applicants’ posts and route “derogatory” determinations into enforcement and visa-related decisions. That follows earlier reporting on ICE’s extensive use of social-media-fed data in immigration enforcement.
The disclosure also places Giant Oak alongside a growing market for government social-media monitoring, after reporting on ShadowDragon’s collection of social data across online platforms for multiple federal agencies. The key issue is not merely collection, but interpretation of speech as an enforcement signal.
First-order effects
- Visa applicants’ social-media posts could be classified by GOST and used as an input to visa and immigration-enforcement decisions, giving ICE a tool-mediated basis for escalating scrutiny.
- Giant Oak’s monitoring service gains a documented operational role across ICE’s enforcement process, while FOIA disclosures expose the system’s use to public and legal scrutiny.
Second-order effects
- Agencies and contractors using comparable social-media-monitoring tools face sharper questions about classification criteria, human review, and whether political or contextual speech is being misread.
- For applicants, the practical risk is that ordinary online expression becomes part of an opaque administrative record, increasing the value of disclosure and challenge mechanisms around automated or tool-assisted assessments.
Third-order effects
- If such tools become routine decision inputs, immigration enforcement shifts from collecting public data toward institutionalized scoring and interpretation of speech—an expansion of the AI enforcement surface that is difficult to audit without records access.
- The durable governance divide will be whether agencies can explain how social signals affect individual outcomes, rather than simply asserting that the underlying posts were publicly available.
The trend: This is one data point in the expansion of government social-data tools from investigative leads into inputs that can shape administrative enforcement decisions.