Media Must Expose ICE's AI-Powered Secrecy Campaign

ICE has quietly contracted with ZeroFox — an AI firm with a documented history of labeling peaceful Black Lives Matter organizers as "threat actors" — to scrub agents' personal information from the internet. The $15 million program, called the "Doxing Mitigation Initiative," was announced in an internal memo leaked to The Intercept. ICE employees who spoke anonymously were blunt: they worry the same technology will be turned inward to identify whistleblowers and critics within the agency itself.

That concern is not paranoid — it is well-founded. ZeroFox's own technology works by seeding information and monitoring who shares it, a technique cybersecurity experts describe as a "honey pot" for catching leakers. This is the same agency that has subpoenaed Reddit for the identity of a user who criticized ICE, that has tried to flip protesters into informants, and that operates primarily out of unmarked vehicles with masked agents. An agency this committed to concealing its own conduct has no business asking taxpayers to fund its anonymity.

We call on the media to cover ICE's ZeroFox contract not as a routine cybersecurity procurement story, but as what it is: a federally funded operation to hide an already unaccountable agency from the public eye — and potentially silence the employees brave enough to report what they see from the inside. The public deserves reporting that asks who ICE is hiding from, and why.