Congress must prevent AI surveillance. The Anthropic feud proves it

March 9, 2026

TL;DR

The US military aims to leverage advanced AI for enhanced surveillance of Americans' movements, search history, and associations.
A dispute arose when the Pentagon found Anthropic's safety guardrails, including no mass domestic surveillance, unacceptable, labeling the company a 'supply chain risk'.
Existing laws are insufficient to address AI's capabilities in tracking, data analysis, and the creation of comprehensive domestic dossiers.
The executive branch can secretly deem surveillance programs 'lawful' without clear congressional rules, raising concerns about warrantless searches of purchased private data.
The Pentagon wants to use AI to analyze unclassified, bulk commercial data on Americans, such as geolocation and web browsing data.
AI tools can rapidly integrate disparate data sources to create detailed profiles of individuals, a capability also sought by agencies like ICE.
OpenAI has amended its Pentagon deal to include language protecting civil liberties, but loopholes remain, and reliance on corporate whims is insufficient.
Congress is urged to pass the Fourth Amendment Is Not For Sale Act and establish basic guardrails on government AI use to protect privacy and freedoms.

Continue reading the original article