Antarbur detection reveals models for Amnesty International for US National Security Customers

Anthropier says A new set of artificial intelligence models Designed for American national security customers.

The new models, a dedicated group of “Claude Gov” models, wrote, “based on direct reactions from our government agents to meet the operational needs in the real world”, as Antarubor writes in the blog post. Compared to models that focus on consumers and institutions, the new GOV models are designed to apply to government operations such as strategic planning, operational support and intelligence analysis.

(These) models are already published by agencies at the highest level of US national security, and access to these models is limited to those working in such classified environments, “Antarbur writes in its post. (They are) they underwent a strict safety test like all our Claude models. “

The Antarbur has increasingly shared US government agents as he is looking for new reliable sources. In NovemberThe company cooperated with Palantir and AWS, the cloud computing department of the main partner and the main investor of Anthropic, Amazon, to sell Amnesty International to defend defense agents.

Antarbur says that its new allocated GOV models deal better with classified materials, “rejects less” when dealing with the classified information, and has a greater understanding of documents in intelligence and defense contexts. The models also have “improved efficiency” in the languages ​​and the important dialects of national security, as Antarbur says, in addition to “improving the understanding and interpretation of the complex cybersecurity data for analysis of intelligence.”

Anthropor is not the best laboratory of artificial intelligence after defense contracts.

Openai is Pursuit To create a closer relationship with the US Department of Defense, and Meta recently open It makes it Lama The models available for defense partners. Google is Refining A copy of Gemini Ai is able to work within the classified environments. During, cohereWhich is built by Amnesty International Company for companies in the first place, also cooperating with Palantir to publish artificial intelligence models, TECHCRUNCH was reported exclusively in early December.