A China-aligned threat group known as TA419 has been attributed by Proofpoint to credential phishing campaigns aimed at AI policy experts in the United States, including personnel at think tanks, universities, and law firms.
The campaigns impersonated high-profile economists and AI policymakers, and in February 2026 targeted an AI policy expert at a U.S. think tank with an outreach concerning the “Request for Feedback on Military Integration of Claude.” Proofpoint notes TA419 has a track record of similar credential-phishing activity against U.S. and Japan-based organisations since at least April 2025, with activity around July 2026 involving impersonations of a former White House Office of Science and Technology Policy leader.
The objective appears to be to gain access to discussions and developments in the U.S. AI policy and regulatory space amid broader geopolitical competition.
The attack sequence begins with seemingly harmless outreach to establish trust, followed by a shortened URL that triggers a multi-stage redirection chain. The final page is an OneDrive-based adversary-in-the-middle (AitM) credential-phishing page that, after a Cloudflare Turnstile check, captures credentials via a Frameless BitB technique—an advanced browser-in-the-browser form of phishing that visually spoofs legitimate pages without using an iframe.
TA419 reportedly enhanced the open-source Frameless BitB tool with telemetry and automation to monitor the target’s Microsoft sign-in flow and relay credentials to real Microsoft infrastructure. The guidance from Proofpoint emphasises the importance of phishing-resistant authentication, such as passkeys, and cautions targets to verify unsolicited outreach before responding.
TA419’s focus has consistently included national security, defence, energy, and policy targets, with the AI policy space representing a natural expansion of that remit.