China-Linked Phishers Target AI Policy Experts

The Record · High sophistication
Last updated October 6, 2026

Researchers reported two China-aligned campaigns that used phishing and impersonation to target AI policy experts and multiple Asian government organizations. One campaign built rapport with “AI policy” themed outreach before sending links to a OneDrive credential-harvesting page, while another used spear-phishing emails with decoy documents to deliver malware for espionage.

How the attack worked

Researchers described two separate China-aligned campaigns targeting people and organizations connected to AI policy and government work. In one, a threat actor impersonated prominent economists and a former White House science and technology policy official to reach AI experts at universities, think tanks, and law firms. The outreach began with a benign conversation-starter email inviting the target to join a fake 'AI Policy Advisory Committee' or contribute to a fictitious Senate report on AI export controls. Only after the victim responded did the attacker send a URL redirection chain that ultimately led to a OneDrive credential phishing page designed to harvest login information.

A second, separate campaign targeted government organizations across multiple Asian countries. This operation relied on phishing emails paired with decoy documents, including news reports, spoofed event invitations, and legislative documents, to get victims to click links or download malicious files. Researchers linked this activity to espionage-focused intelligence gathering across a notable number of compromised endpoints spanning several countries.

Why it succeeded

Both campaigns relied on patient, topic-relevant social engineering rather than obvious red flags:

  • The first campaign used rapport-building before ever sending a malicious link, so the initial message looked harmless and professionally credible.
  • Impersonating recognizable names in the AI policy space added legitimacy that made the invitation plausible to busy researchers and policy staff.
  • The second campaign leaned on timely, topical content (legislative updates, real event invitations) that targets would be inclined to open without hesitation.
  • Credential theft occurred only after a redirect chain, separating the initial contact from the actual malicious action and reducing suspicion at each step.

What to watch for

  • Unsolicited invitations to join advisory committees, contribute to reports, or collaborate on policy work, especially from contacts you were not expecting to hear from.
  • Early emails that seem purely conversational, with no attachment or link, since these may be designed only to solicit a reply before a follow-up attack.
  • Links that redirect multiple times before landing on a login page, particularly one resembling OneDrive or another cloud service.
  • Attachments or links framed around breaking news, legislative updates, or event invitations that arrive unexpectedly.

How to build resistance

Organizations supporting policy, research, or government functions should treat unsolicited collaboration requests as high-risk even when they appear to come from credible experts. Staff should be trained to pause before replying to conversation-starter emails from unfamiliar senders and to verify legitimacy through known official channels rather than the contact details provided in the email itself. Credentials should never be entered after following a link from an email, especially after being redirected to a cloud login page. Because these campaigns used topical, well-researched lures, awareness training should emphasize that realistic and relevant-seeming content is not the same as verified content.

Key findings

  • Proofpoint described a July incident where a China-aligned actor ran phishing attacks by impersonating prominent economists and a former White House science/technology policy official.
  • Targets included AI experts in universities, think tanks, and law firms, with lures about joining a fake “AI Policy Advisory Committee” or contributing to a fictitious Senate report.
  • After victims responded, attackers sent a URL redirection chain leading to a OneDrive credential phishing page.
  • Cisco Talos reported a separate China-linked campaign targeting government organizations across multiple Asian countries, with initial access via phishing emails and decoy documents.
  • Talos observed ~350 compromised endpoints across eight countries and stated the campaign goal was intelligence gathering.

Who’s being targeted

  • Commonly targeted roles: Executive leadership, Government affairs / public policy, Researchers (AI/technology), Think tank staff, Legal (policy/regulatory), Administrative staff supporting policy teams.
  • Affected industries: Government, Higher education, Think tanks / policy organizations, Legal services, Artificial intelligence / technology.
  • Attack channels: email, website.
  • Impersonated: Former White House official Lynne Edwards Parker / foreign policy expert Heidi Crebo-Rediker (impersonation), Unspecified (campaign used decoy documents and spoofed invitations).

Red flags to watch for

  • Too-good-to-be-true policy invitation from an unexpected sender
  • Conversation-starter email designed to prompt a reply before any real details are provided
  • Link redirects and leads to a login page asking for credentials (OneDrive phishing)
  • Unexpected email with attachments/decoy documents
  • Pressure to review timely ‘news’ or ‘legislative’ material
  • Sender/event details don’t match known official channels
Try Mirage

Mirage safely runs attacks like this one against your own team, so you find out what happens before a real adversary does.

Get a demo

Frequently asked questions

What pretext did attackers use to target AI policy experts?

They sent benign conversation-starter emails inviting targets to join a fake 'AI Policy Advisory Committee' or contribute to a fictitious Senate report on AI export controls, building rapport before sending malicious links.

What happened after a victim replied to the phishing email?

The attackers followed up with a URL redirection chain that led to a OneDrive credential phishing page designed to steal login information.

Who did the second campaign target and how?

A separate China-linked campaign targeted government organizations across multiple Asian countries using phishing emails and decoy documents, including news reports, spoofed event invitations, and legislative documents, to deliver malware for intelligence gathering.

Why were these campaigns considered sophisticated?

They combined impersonation of credible experts, multi-stage rapport building, and topical lures tailored to the target audience, which researchers assessed as highly sophisticated social engineering.

Read the video transcript

You get this email: “Invitation to join an AI Policy Advisory Committee,” supposedly from a former White House science official. Researchers say China-linked groups are doing this for real, impersonating economists and policy experts, asking you to join a fake AI committee or contribute to a Senate report, just to get you to reply. Here’s the trick: once you respond, they send a link that bounces through redirects into a OneDrive login page. It looks normal, but it’s a credential phish built to steal your username and password. If you get an unexpected policy invite or briefing, and it leads to a login page after a few redirects, stop. Don’t sign in there, go to the service directly and log in from a fresh browser tab.

Similar attacks

Fake Recruiters Target Job Seekers With Malware

Fake Recruiters Target Job Seekers With Malware

Iran-aligned DarkBlinders used convincing recruitment offers while posing as Dubai Airports recruiters to target a high-value job seeker in Iraq. The victim was asked to complete what appeared to be a coding assessment, but the assessment package installed malware. The group also used lookalike…

November 2, 2026
North Korea Scales AI-Powered Job Scams

North Korea Scales AI-Powered Job Scams

Microsoft reports that North Korean threat actors are using AI to make job scams, fake recruiter outreach, fraudulent worker identities, and targeted phishing campaigns faster and more convincing. One group posed as recruiters for a spoofed AI blockchain company and scheduled software-engineer…

October 29, 2026
Device-Code Phishing and “ClickFix” Lures Spread

Device-Code Phishing and “ClickFix” Lures Spread

This weekly recap highlights multiple real-world campaigns where attackers trick users into taking actions that grant access, without needing to steal passwords directly. Notable examples include “device code” phishing (victims are instructed to enter a short code to approve an attacker session)…

September 28, 2026
Fake Gmail Attachment Lure Drops Antino Backdoor

Fake Gmail Attachment Lure Drops Antino Backdoor

A China-nexus threat group targeted government and policy organizations across Asia using spear-phishing emails tailored to the victim’s interests. The emails used spoofed trusted senders and a realistic fake Gmail attachment preview that linked to attacker-controlled pages, ultimately installing…

October 2, 2026
Fake Google Play Pages Push Spyware at Logistics

Fake Google Play Pages Push Spyware at Logistics

A real campaign is targeting logistics firms with fake Google Play pages impersonating well-known logistics brands to trick employees into installing an Android spyware app. Once installed, the spyware can steal newly received SMS messages (including one-time passcodes) and enable call forwarding,…

September 24, 2026
China-Linked Phish Uses Fake Gmail Preview

China-Linked Phish Uses Fake Gmail Preview

A China-linked espionage group (UAT-11587) targeted Asian government and policy organizations using highly tailored phishing emails and realistic decoy documents. After a click, malware ultimately installed the “Antino” backdoor, which then hid its command-and-control traffic inside normal…

October 3, 2026