A $700K sales job points Anthropic back to the Pentagon

Anthropic is hiring a Head of National Security Sales at a salary of up to $700,000 a year, according to a public job posting reviewed by Prospect. The listing, which had not been previously reported, appeared this week, months after the Pentagon cut the company off from federal contracting.

At a glance

• The role calls for building out a sales organization aimed at the Department of War and the Intelligence Community, with the stated goal of driving adoption of safe, frontier AI across both.
• The Pentagon labelled Anthropic a national security supply chain risk, a designation that barred it permanently from contracts with the federal government or any federal contractor.
• A federal appeals panel is expected to rule imminently on whether the Defense Department broke the law in blacklisting Anthropic, a ruling that would decide whether the company can bid for defense work again.

The hire reads as preparation rather than resolution: Anthropic is staffing a function that the current designation leaves it no contracts to serve, which suggests the company expects the ban to fall. It also puts weight on the red lines the company has made central to its public identity, since military and intelligence buyers are pushing hardest on exactly the surveillance and autonomy applications Anthropic said it would refuse.

The role targets the Department of War and the Intelligence Community while the blacklist stands

Per the description, the new head of national security sales would take a leadership role in rebuilding Anthropic's contracting relationship with military and intelligence agencies. The hire is expected to "lead and scale our national security sales organization to drive the adoption of safe, frontier AI across the Department of War and the Intelligence Community."

The Defense Department designated Anthropic a national security "supply chain risk" earlier this year, a move that summarily cut it off from federal contracting and from work with federal contractors. Anthropic has contested the designation for months, arguing it was politically motivated retaliation over the red lines it drew on mass domestic surveillance or fully autonomous weapons.

The White House called Anthropic a "radical left, woke company" before the Pentagon signed OpenAI

The designation followed the disclosure that Claude had been used in the military raid to capture Venezuelan President Nicolas Maduro, and came shortly before Anthropic's technology was implicated in a US drone strike on an Iranian girls' school. The White House called the company a "radical left, woke company" run by "leftwing nut jobs."

Hours later the Defense Department signed an agreement with OpenAI, which said its services would not be used by Department of War intelligence agencies like the NSA and would not be intentionally used for domestic surveillance of US persons and nationals. OpenAI's head of robotics, Caitlin Kalinowski, resigned over the deal.

Surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got.

Brennan Center counsel Amos Toh says AI can surface thousands of target recommendations at once

Amos Toh, senior counsel at the Brennan Center's Liberty and National Security Program, said automation and AI have raised the pace of warfare, and that systems able to identify hundreds or thousands of target recommendations mean the scale of harm will be greatly amplified. Models can also produce recommendations that sound extremely convincing while being thoroughly off base, he said.

In Ukraine this month, a fully autonomous Russian drone killed three civilians as it descended on a gas station. Toh also pointed to agent failures off the battlefield, including OpenAI's hack of a competitor's data systems this summer, which went unnoticed for more than a week, and to the possibility that an agent leaks classified information without the military or the developer knowing.

What the appeals ruling settles

A federal appeals panel is set to rule imminently on whether the Defense Department broke the law in blacklisting Anthropic, and no date for the decision has been given. Whether restoring the relationship requires concessions to military and intelligence community demands is not stated in the posting.

OpenAI is also expanding its national security contracting team, hiring sales representatives with top secret clearance to move defense and intelligence agencies from consideration to deployment, and its Pentagon agreement left open the possibility of new deals with agencies like the NSA.