The White House Reaches Agreement with Anthropic on Intelligence Agencies' Use of AI

According to The New York Times, the White House is in the final stages of negotiating an agreement with AI company Anthropic, which would allow the US National Security Agency (NSA) and other intelligence agencies to use Anthemp's advanced AI models for classified work.

Reversal under the Shadow of "Supply Chain Risks"

The context of these negotiation discussions is quite delicate. Earlier this year, the US Department of Defense labeled Anthropic as a "supply chain risk" company – the trigger was a dispute over existing contract terms: the Pentagon wanted to use Anthropic technology for mass surveillance of American citizens or autonomous weapon combat, and Anthropic raised objections to this. Currently, the company is suing the Department of Defense over the aforementioned designation.

The report suggests that the terms of the proposed secret work contract will address Anthropic's core concerns, opening a new window for the relationship between the two parties.

The Mythos Model as a Key Bargaining Chip

Anthropic's bargaining power in the negotiations has increased significantly in recent months, following the company's disclosure of a new generation AI model named Mythos, which has performed particularly well in identifying software vulnerabilities, aligning with the core needs of intelligence agencies. US national security and military institutions have been using Anthropic models in actual work for a long time and are seeking to maintain this cooperative relationship under the pressure of the "supply chain risk" designation.

The $9 Billion Chip Procurement Plan Moves Forward

The report also reveals that the White House has approved a $9 billion procurement application for Nvidia Blackwell chips proposed by intelligence agencies to provide computing power support for the aforementioned AI applications. This substantial hardware investment further highlights the US intelligence community's determination to prioritize the construction of AI capabilities.

Should these negotiations ultimately come to fruition, it will mark a significant reconfiguration of the relationship between Silicon Valley AI companies and US national security agencies – Anthropic is attempting to forge a path that balances business interests with ethical baselines amidst compliance boundaries and national strategic needs.