Anthropic claims to have unveiled a fresh batch of AI models designed specifically for national security clients in the United States.
According to Anthropic’s blog post, the new models—a customized set of “Claude Gov” models—were “developed based on direct feedback from our government customers to solve real-world operational demands.” The new bespoke Claude Gov models were intended to be utilized for government operations such as intelligence analysis, operational assistance, and strategic planning, in contrast to Anthropic’s consumer- and enterprise-focused models.
“Only those who work in such classified contexts have access to these models, and [these] models are already implemented by agencies at the highest level of U.S. national security,” Anthropic adds in its post. As with all of our Claude models, “[they] passed the same stringent safety testing.”
As it searches for stable new revenue streams, Anthropic has been working with U.S. government clients more and more. In order to offer Anthropic’s AI to defense clients, the business partnered with Palantir and AWS, the cloud computing branch of Anthropic’s main investor and partner, Amazon, in November.
According to Anthropic, their new custom Claude Gov models have a higher grasp of documents in intelligence and defense contexts, handle classified material better, and “refuse less” when interacting with secret information. Additionally, according to Anthropic, the models have “increased understanding and interpretation of complicated cybersecurity data for intelligence analysis” and “enhanced competency” in languages and dialects essential to national security operations.
Anthropic isn’t the only top AI lab going after defense contracts.
Meta recently announced that it is making its Llama models available to defense partners, and OpenAI is working to forge a stronger collaboration with the U.S. Defense Department. Google is developing a Gemini AI version that can operate in classified settings. Meanwhile, TechCrunch exclusively revealed early last December that Cohere, which mostly develops AI solutions for companies, is working with Palantir to implement its AI models.
Leave feedback about this