U.S. Media: Despite Potential Risks, NSA Continues Using Anthropic's Most Powerful AI Model
Axios website cited sources reporting that, despite the Pentagon stating that Anthropic’s artificial intelligence (AI) technology poses risks, the U.S. National Security Agency (NSA) continues to use the company’s most powerful AI model—Mythos Preview.
The site noted: "The NSA is using Anthropic’s currently strongest model, Mythos Preview, despite senior Defense Department officials responsible for overseeing the NSA insisting that the company poses a risk to the supply chain."
It remains unclear exactly how the NSA is utilizing the Mythos Preview model, but other organizations with access primarily use it to scan their own systems in search of potential security vulnerabilities that could be exploited by hackers.
Previously, Anthropic has restricted access to Mythos for approximately 40 organizations, citing that the technology’s network capabilities are too dangerous to be widely disseminated.
Earlier, the Pentagon demanded that Anthropic grant unrestricted military access to its Claude AI technology for use in military operations, threatening to cancel contracts and disrupt the supply chain if Anthropic refused. This dispute over AI technology access aligns with Anthropic’s position: the company’s leadership has declined to provide full access to the Pentagon. Anthropic believes it is still premature to rely on these technologies to develop autonomous weapons.
Anthropic also maintains that there is currently no legal basis for using its technology in domestic surveillance.
Source: sputniknews
Original article: toutiao.com/article/1862963685289996/
Disclaimer: The views expressed in this article are those of the author(s) alone.