Executive Summary
Since at least April 2025, a China-aligned espionage group designated as TA419 has targeted individuals at U.S. and Japanese think tanks, universities, and law firms. Throughout 2026, the actor shifted focus toward AI policy experts, impersonating high-profile figures from the White House Office of Science and Technology Policy and AI safety organizations like Anthropic.
The attack chain involves social engineering via "benign" outreach regarding AI regulation, followed by multi-stage URL redirections. Victims are led to an Adversary-in-the-Middle (AitM) phishing page that utilizes "Frameless BitB," a sophisticated Browser-in-the-Browser technique that lacks traditional iframe markers. This kit captures credentials and MFA session cookies in real-time by relaying the authentication flow to legitimate Microsoft infrastructure.
This activity represents a strategic effort by Chinese intelligence to gain insights into U.S. AI policy, regulatory developments, and military integration of AI models. The technical sophistication of the AitM kit allows the actor to bypass standard multi-factor authentication, posing a significant risk to high-value targets in the AI research and policy space.