GrapheneOS is moving on from Pixels? ThreatWire

In the latest cyber security news roundup for the week of March 2, 2026, host Allie Diamond highlights significant developments in mobile operating systems and a high-stakes conflict between AI giant Anthropic and the U.S. government. The episode begins with the announcement of a partnership between Motorola and GrapheneOS. Known for its rigorous security and privacy standards, GrapheneOS was previously exclusive to Pixel phones. This collaboration aims to bring the secure, Android-compatible operating system to Motorola hardware, provided the devices meet strict hardware-based security requirements. This move reflects a growing trend in mobile security, following Apple’s recent success in achieving NATO approval for iPhones and iPads to handle classified information without third-party modifications. The focus then shifts to Anthropic, which has experienced a turbulent week. On February 20, the company launched "Cloud Code Security," an AI-driven tool designed for vulnerability scanning and patch suggestion. However, this launch was quickly followed by reports of "distillation attacks." Anthropic discovered that Chinese AI groups, including Deepseek and Moonshot, used thousands of fraudulent accounts to train their proprietary models using Claude’s outputs. Anthropic warned that such attacks allow foreign entities to bypass American export controls, potentially feeding advanced AI capabilities into military and surveillance systems. Compounding these issues, Anthropic faced internal security challenges when three vulnerabilities were discovered in Cloud Code, leading to risks of remote code execution and credential theft. More significantly, the company is currently in a standoff with the U.S. Department of Defense (DoD). Anthropic’s CEO revealed that the government is pressuring the company to allow Claude to be used for mass domestic surveillance and autonomous weaponry. Anthropic has refused, citing ethical concerns and the belief that current AI models are not suited for such tasks. In response, the Secretary of Defense has threatened to label Anthropic a "supply chain risk"—a designation usually reserved for foreign adversaries. This label would effectively bar any military contractor from doing business with the company. In the wake of this conflict, OpenAI has stepped in to fill the gap. CEO Sam Altman announced that OpenAI would collaborate with the U.S. government, asserting that their deployments include stricter guardrails than Anthropic’s and will not be used for surveillance. The episode concludes with a brief discussion on password security. Diamond reflects on the inherent vulnerabilities of browser-first password managers, suggesting that online-only solutions remain a significant point of concern for security practitioners.