AI-powered cognitive warfare conducted by China and Russia is shifting from mass influence operations to highly personalized efforts tailored to the interests, emotions, and vulnerabilities of individuals.
People must be warned in advance about common tactics such as impersonation accounts, fake digital influencers, manipulated videos, and AI personas so that they can recognize when something feels wrong.
Technology platforms bear a grave responsibility to detect and remove fraudulent identities and deceptive content.

Why I Am Republishing This Article

I am republishing this article for the operator of the account impersonating my former pen name, “Kenji Akutagawa,” and for Meta, which has continued to leave that account in place despite repeated reports.
An impersonation account is not merely a prank, nor is it a private nuisance that the person being impersonated should simply be expected to endure.
It is a means of manipulating information by using another person’s name, photographs, background, or writings to create a fraudulent identity and mislead readers, friends, acquaintances, and the wider public.
Whether a particular impersonation account was created with AI or is operated manually by a human being is not the essential issue.
The essential issue is the creation of a false identity that pretends to be a real person, appropriates that person’s credibility and social relationships, and attempts to influence the judgment of others.
The article by Vanderbilt University professor Brett Goldstein, published yesterday in the Opinion section of the Sankei Shimbun, explains that impersonation accounts, fake videos, synthetic influencers, and AI personas have already become practical instruments of state-sponsored cognitive warfare.
Editorial writer Ken Sato then identifies the most important countermeasure in one sentence:
What is needed is to warn people in advance about typical tactics such as impersonation accounts and fake videos, so that recipients can recognize when something feels wrong.
That is precisely what I have demanded from Meta for years.
Impersonation must not be dismissed as the victim’s private problem.
It must be investigated as information manipulation carried out through a fraudulent identity, removed, and prevented from recurring.
The following is from an article published in the Opinion section of the Sankei Shimbun yesterday, August 1, 2026.
China Uses AI Cognitive Warfare to Target Individuals
Brett Goldstein
Professor, Vanderbilt University
The methods used in influence operations known as “cognitive warfare” are undergoing a major evolution.
We are moving away from an era of manual operations and into an era in which such activities are strengthened, streamlined, and driven by artificial intelligence.
Influence operations conducted through social media are not themselves new.
During the 2016 and 2020 presidential elections in the United States, Russia conducted operations intended to polarize public opinion.
China has also sought to inflame American public opinion.
During the 2024 presidential election, operators posed as Americans and attempted to deepen social divisions over highly contentious issues and United States foreign policy.
Compared with the technology-driven methods now emerging, however, those techniques were primitive.
In August 2025, my Vanderbilt University research partner, Mr. Benson, and I published our findings concerning a Chinese company called GoLaxy.
The company had developed AI tools for influence operations targeting Japan, Hong Kong, and Taiwan, with Southeast Asia and the United States included as additional targets.
It collects publicly available information from across social media and develops hundreds of “personas”—virtual identities trained to appear as human as possible—which distribute messages designed to change people’s behavior.
According to the company’s own materials, it performed this work under commission from the Chinese government and intelligence agencies.
Such AI personas have already been deployed in Taiwan, and foundational stages of similar operations are now being observed in the United States and Japan.
These are the first signs that AI-powered influence operations at the individual level are already taking place, and the situation is likely to worsen.
AI shifts foreign influence operations from the group level to the individual level, enabling a scale and sophistication that were previously impossible.
To one person, an AI persona may appear to be a celebrity.
To another, it may appear to be someone who shares the same interests, or an entertaining and provocative influencer.
They approach their targets and behave in a friendly manner, but they act under the direction of a nation-state and gradually implant ideas intended to guide the individual in a desired direction.
AI and advanced models should continue to evolve.
We must, however, honestly confront the darkness that accompanies AI.
Fake digital influencers, manipulated videos, and attacks on elections are already spreading throughout our daily lives.
Unless we act, it will become increasingly difficult to trust anything—or anyone—we encounter online.
Everyone has a role to play in identifying these personas and reducing their potential influence.
Individuals must first know that these tools exist.
There are, of course, limits to what individuals can do.
The academic community must cooperate and discover new ways to identify synthetic messages in an increasingly disorderly world.
Governments and leading AI laboratories bear the greatest responsibility.
They cannot fulfill that responsibility without cooperation.
The challenge is not to slow technological progress, but to make it possible to operate safely even at full speed.
Japan and its allies should define AI-powered influence operations as a shared threat and respond to them in the same coordinated manner used to confront cyberattacks.
Japan Is Being Targeted—Preparations Must Be Accelerated
With the evolution of AI, cognitive warfare conducted by China and Russia is moving beyond ordinary propaganda and toward techniques tailored to the interests of individual people.
“AI personas” that behave like human beings may approach a target while pretending to be a friend or an influencer and then attempt to affect that person’s judgment.
Japan must recognize that it is being targeted.
Simply warning people to “watch out for personas” is not enough.
Chinese and Russian influence operations are not so simple that they can always be identified easily.
What is needed is to warn people in advance about typical tactics such as impersonation accounts and fake videos, so that recipients can recognize when something feels wrong.
Governments and AI companies must cooperate and improve the technologies used to detect fraudulent communications.
What is at stake is our ability to protect the democratic information environment from influence operations that exploit AI.
Ken Sato
Editorial Writer

What Meta Must Do

This article makes clear the danger of leaving impersonation accounts untreated as though they were merely private disputes between individuals.
Impersonation persuades third parties to believe in a fabricated identity, distorts their understanding of the real person, and appropriates that person’s reputation, work, and social relationships.
Meta must immediately investigate the account using my former pen name, “Kenji Akutagawa,” including its identity verification, creation and operating history, related accounts, and unauthorized use of photographs and written material.
Meta must also remove the account, prevent the same operator from recreating it, and explain why repeated reports were left unresolved for so many years.
Public education that helps recipients recognize warning signs and swift detection and removal by the platform are both indispensable.
Without both, information manipulation through impersonation cannot be stopped.