China-Linked Espionage Group Sends Phishing Campaign to U.S. AI Policy Makers

A cyber‑espionage group with ties to China has launched a targeted phishing operation against individuals involved in shaping artificial intelligence policy in the United States, according to a report released Thursday by security firm Proofpoint.

The campaign relies on deceptive emails that masquerade as messages from high‑profile government officials, noted economists and even a staff member of the AI research firm Anthropic. Recipients are prompted to click malicious links or open attachments that can install malware or harvest credentials, giving the attackers a foothold in networks that influence AI regulation and strategy.

Proofpoint’s analysis indicates the operation has been active for several months, focusing on a narrow pool of experts who advise legislators, draft policy proposals, or work for think‑tanks focused on emerging technologies. By impersonating trusted figures, the attackers increase the likelihood that recipients will comply, a technique that mirrors broader trends in state‑sponsored cyber activity where social engineering is combined with technical exploits.

While the specific group has not been publicly named, its methods and infrastructure align with previous campaigns attributed to Chinese intelligence services. The use of an Anthropic employee’s identity is notable, reflecting the growing interest of foreign actors in the private‑sector AI ecosystem, where proprietary models and research can yield strategic advantage.

U.S. officials have repeatedly warned that adversaries are seeking to influence the development and governance of AI, a technology considered a cornerstone of future economic and military power. The phishing attempts underscore the vulnerability of policy networks that, unlike typical corporate IT environments, may lack robust security training and resources.

Industry observers say the incident highlights a need for heightened awareness among policymakers and researchers alike. Best practices such as multi‑factor authentication, email verification tools and regular phishing simulations are being recommended to mitigate the risk of credential theft and malware infection.

Proofpoint’s findings arrive as Washington debates a suite of AI‑related bills aimed at ensuring transparency, accountability and national security. If successful, the espionage effort could give the originating nation insight into U.S. legislative priorities and the technical underpinnings of emerging AI systems, potentially shaping the global balance of AI leadership.