Chinese Hackers Impersonate AI Experts to Target US Policy Researchers

A China-aligned hacking group has been impersonating well-known artificial intelligence experts and officials to target US policy researchers, according to cybersecurity company Proofpoint.

The group, identified as TA419, has been active since April 2025. Proofpoint said it targeted AI policy experts in July by pretending to be prominent figures in the technology and government sectors.

One impersonated person was reportedly Lynne Edwards Parker, a former senior official at the White House Office of Science and Technology Policy.

Fake AI policy invitations used as bait

The attackers initially sent emails that appeared harmless.

In some cases, targets were invited to join an “AI Policy Advisory Committee” or participate in other AI-related discussions.

The emails were designed to start a conversation rather than immediately steal information.

Once a target responded, the attackers sent them to a fake login page designed to collect their account credentials.

The targets included policy experts working at think tanks, defence contractors, universities and law firms in the United States and Japan.

Fake login window makes phishing harder to spot

According to Proofpoint, the attackers used a technique known as a browser-in-the-browser attack.

Instead of sending victims to an obviously suspicious website, the attackers created a fake browser pop-up inside a legitimate webpage.

The fake window was designed to look like a genuine login prompt, making it more difficult for victims to realise that they were entering their credentials into a phishing page.

This type of attack takes advantage of something users already trust: the appearance of a familiar login screen.

Former White House official among impersonation targets

Proofpoint did not publicly identify all the people targeted.

However, Reuters confirmed that Alex Engler, a former White House official who now leads the Penn Center on Media, Technology, and Democracy, received one of the emails.

Engler reportedly checked the message with colleagues in the AI industry and discovered that it had been sent by someone impersonating an expert.

The incident shows how attackers can use trusted professional identities to make targeted phishing emails appear more credible.

AI policy experts increasingly targeted

TA419 has reportedly used similar tactics before.

In February, the group was linked to an attempt to impersonate a prominent employee of AI company Anthropic while targeting AI policy experts.

Proofpoint believes the group is likely to continue targeting think tanks and other organisations involved in AI policy.

The use of real people's identities is particularly effective in highly specialised industries, where professionals regularly communicate with researchers, government officials and industry experts they may not personally know.

Why the attack matters

The campaign demonstrates how modern phishing attacks are becoming more targeted.

Instead of sending thousands of generic emails, attackers can research a person's professional interests and create messages that match their work.

For organisations working in AI policy, technology and national security, this can create additional risks because employees may regularly receive invitations to conferences, advisory groups, research projects and policy discussions.

Security teams can reduce the risk by verifying unexpected invitations through a separate communication channel, checking the sender's email address carefully and avoiding login links included in unsolicited messages.

Using multi-factor authentication can also limit the damage if a password is stolen.

The campaign highlights a simple but important cybersecurity lesson: a familiar name or professional invitation does not necessarily mean the person behind the message is genuine.