A China-aligned cyber espionage actor tracked as TA419 has targeted artificial intelligence policy experts across U.S. think tanks, universities, and legal organizations, security firm Proofpoint reported. The group's credential phishing operations aim to capture sensitive access, likely in support of broader intelligence gathering around U.S. AI regulatory frameworks, trade policy, and model safety developments.
Impersonation tactics and targets
Proofpoint noted that TA419 has conducted targeted phishing since at least April 2025, historically focusing on defense, energy, foreign policy, and national security entities in the U.S. and Japan. In February 2026, the group targeted an AI policy expert at a U.S. think tank using emails impersonating prominent AI policymakers, economists, and an Anthropic employee, under subject lines such as "Request for Feedback on Military Integration of Claude."
By July 2026, the adversary expanded its impersonation tactics to include a former leadership member of the White House Office of Science and Technology Policy. The campaign begins with benign outreach to build trust before delivering malicious links to victims who respond.
Frameless BitB and AitM technical flow
When a target engages, TA419 delivers a shortened URL that initiates a multi-stage redirection sequence. After passing a Cloudflare Turnstile challenge, the link loads an adversary-in-the-middle (AitM) phishing page hosted via OneDrive.
To deceive targets, the phishing infrastructure uses a technique known as Frameless Browser-in-the-Browser (BitB). Unlike traditional BitB attacks that rely on HTML iframes, Frameless BitB uses custom scripts, HTML, and CSS injections to simulate a legitimate login interface directly within the browser session. TA419 integrated a custom telemetry and automation module into the open-source tool, allowing it to proxy Microsoft sign-in flows in real time. The proxy captures user credentials and session cookies while successfully completing the authentication flow against Microsoft's actual servers, leaving the victim unaware of the compromise.
Defensive recommendations
Proofpoint stated that targeting AI policy figures extends TA419's existing focus on strategic national security and technology topics rather than representing a new direction. To mitigate the threat, security researchers advise organizations to implement phishing-resistant multi-factor authentication methods, such as passkeys, and urge experts to independently verify unsolicited subject-matter outreach before opening links.
AI-generated editorial illustration; not a documentary photograph.
Source: China-Aligned TA419 Targets U.S. AI Policy Experts With Microsoft AitM Phishing
SIGNAL TITANS
Leave a Reply