Security company Proofpoint said Thursday that a phishing campaign targeted AI policy experts at U.S. think tanks, universities and law firms with invitations apparently sent by prominent policy figures. The campaign sought access to cloud accounts, but Proofpoint did not report whether anyone’s account was compromised.
Proofpoint attributed the activity to a group it tracks as TA419, which it assesses to be aligned with China and motivated by espionage. That is the company’s assessment, not an independently established government link. CyberScoop reported that Proofpoint did not directly connect the operation to the Chinese government.
There is independent confirmation of one narrower finding. Lynne Parker, a former senior White House official whose identity was used in the emails, told Nextgov/FCW that two recipients contacted her on July 9 to check whether she had sent the messages. Her account does not establish how many people received them or whether anyone followed a phishing link.
The distinction matters to the researchers and institutions involved: A successful attempt to access a policy expert’s cloud account could expose correspondence and contacts. The available reporting establishes an attempt to obtain that access, not a disclosure of private information.
How the invitations led toward a fake sign-in
Proofpoint said the July campaign began on July 8. The emails impersonated Parker, who formerly served as principal deputy director of the White House Office of Science and Technology Policy, and Heidi Crebo-Rediker, an economist and foreign-policy expert.
The messages invited recipients to join a purported AI policy advisory committee or contribute to a supposed Senate Foreign Relations Committee report on AI export controls and supply chains. Proofpoint described those projects as lures, not genuine invitations from Parker, Crebo-Rediker or the committee.
According to Proofpoint, the initial emails sought a conversation rather than a password. After a recipient replied, a follow-up directed the person toward a fraudulent Microsoft OneDrive sign-in page. The attempt was designed to look like a routine continuation of professional correspondence.
Proofpoint said the final phishing stage was designed to capture sign-in details and information from an active account session. It identified a customized version of an existing phishing tool in the setup. That description explains what the attempt was built to do; it does not show that anyone entered credentials or that the operators obtained a working session.
The company said the attempt targeted accounts associated with Microsoft 365 and Entra ID. Its report did not identify a compromised account or say how many recipients reached the false sign-in. Receiving an invitation, replying to it and losing account access are different outcomes, and the reporting does not establish the last of those for anyone.
Why Proofpoint linked the activity to TA419
Proofpoint said its observation of TA419 dates back to at least April 2025. It has seen the group approach people connected to think tanks in the United States and Japan, along with people at defense contractors, universities and law firms. Proofpoint assessed the July focus on AI policy as consistent with what it describes as the group’s interest in defense, national security and foreign policy.
Free newsletter
Get the morning briefing
Start each day with the AI stories that matter and why — a short, free email from our newsroom.
The company also described a separate February 2026 approach to an AI policy analyst at a U.S. think tank. That effort impersonated a senior Anthropic employee, whom Proofpoint did not name. The reporting does not establish that the February approach or the July campaign resulted in account access.
Those findings are part of Proofpoint’s investigation; another outlet recounting them is not an independent verification of who operated the campaign. The company’s assessment that TA419 is aligned with China does not, by itself, establish Chinese government direction. Nor does the stated focus on policy establish that the operators obtained policy documents.
CyberScoop reported that Chinese officials have repeatedly denied conducting cyberespionage. That general position is not a response to Proofpoint’s specific findings. Nextgov/FCW reported seeking comment from China’s embassy but did not publish a response.
What recipients confirmed, and what remains unknown
Parker said the two people who contacted her on July 9 reached her through separate channels. She told them the messages were fraudulent and alerted colleagues. Their inquiries corroborate that messages using her identity reached at least two people; they do not provide a count of everyone targeted.
Parker’s warning to Nextgov/FCW — “The challenge is that, without knowing who the bad actors are targeting, it’s difficult to reach everyone who might be at risk” — reflects the limits of contacting only known recipients.
Don Styer, a federal-services executive at Consulting Solutions, told Nextgov/FCW that ordinary-looking requests and impersonation can exploit professional trust. He said human judgment remains an important defense. In this case, two recipients’ decision to check directly with Parker provided a warning, but it cannot establish what happened to messages sent to others.
Proofpoint predicts continued targeting and identity spoofing by the group, but that is a forecast, not a confirmed future operation. No subsequent investigation or action is established in the reporting. The central unanswered questions are how many people received the invitations and whether the fake sign-in yielded any account access.