Japan has issued warnings about the growing scale of foreign influence operations after a series of investigations revealed coordinated campaigns linked to China targeting the country’s political leadership and democratic processes.
The revelations, based on reports by technology firms and research institutes, have prompted senior Japanese officials to call for urgent countermeasures, describing such activities as a direct threat to the foundations of democratic governance.
At the centre of the controversy is a report released in February by OpenAI, which disclosed that it had banned a ChatGPT account linked to a Chinese law enforcement actor who attempted to use artificial intelligence tools to plan a covert influence operation against Japanese political figures.
According to the report, the operation included plans to undermine the reputation of Japan’s prime minister, Sanae Takaichi.
The findings triggered immediate concern in Tokyo. Speaking at a regular briefing on January 27, Japanese Chief Cabinet Secretary Minoru Kihara described foreign influence operations as a “national security threat,” warning that such campaigns could undermine core democratic principles, including free elections and a free press.
AI tools and a planned influence campaign
The OpenAI report revealed that the China-linked user attempted to design a multiphase information campaign using the company’s AI system.
The request reportedly appeared in mid-October 2025, shortly after Takaichi publicly criticised Beijing’s policies toward ethnic minorities in the region of Inner Mongolia.
At a public meeting focused on human rights conditions there, Takaichi condemned restrictions imposed on ethnic Mongolian communities.
In recent years, the Chinese Communist Party has expanded Mandarin-only education policies in parts of the region, a move critics say is part of a broader assimilation campaign affecting minority communities.
According to the OpenAI report, the user asked the AI system to craft a plan to discredit Takaichi at a politically sensitive moment, when she was expected to emerge as a leading contender to head Japan’s ruling Liberal Democratic Party.
The draft strategy included amplifying negative online commentary about the Japanese politician, criticising her immigration proposals, and posing as foreign residents sending complaints to Japanese lawmakers.
The plan also suggested using fake accounts to blame Takaichi for rising living costs and stirring public anger over U.S. tariffs on Japanese agriculture in order to divert attention away from China-related issues.
OpenAI said it refused to assist with the request and later identified attempts by the same user to refine a progress report on the influence operation.
Signs of activity on social media
Researchers later identified social media accounts posting coordinated messages about Takaichi.
In late October 2025, accounts were detected using hashtags portraying the prime minister as a far-right figure and raising concerns about economic consequences linked to foreign trade disputes.
By November, some of the same accounts were spreading English-language posts criticising Takaichi’s comments about Taiwan. The activity suggested that operators were targeting both domestic and international audiences in order to shape perceptions of Japan’s leadership.
OpenAI researchers said the activity indicated a well-resourced strategy involving human operators supported by artificial intelligence tools. The user’s requests referenced a broader influence infrastructure described as “cyber special operations,” reportedly staffed by hundreds of individuals.
A wider network of online influence
Parallel investigations have reinforced the scale of the operation. A report released in February by the Washington-based think tank Foundation for Defence of Democracies identified a network of more than 330 social media accounts linked to pro-Beijing narratives targeting political figures in multiple countries.
According to the analysis, dozens of accounts operating on the social media platform X and on Tumblr pushed messages attacking Takaichi during the period surrounding her election victory.
The posts portrayed her as an illegitimate or militaristic leader and circulated corruption allegations.
Researchers concluded that the network was part of a broader influence campaign targeting political developments in Japan, the United States, the Philippines, and several Latin American countries.
The messages frequently echoed pro-China policy positions while criticising governments seen as opposing Beijing’s strategic interests.
Although individual posts often received limited engagement, analysts noted that the accounts appeared designed to manipulate algorithms and increase the visibility of certain narratives. Even messages with minimal interaction could still reach thousands of viewers.
‘Spamouflage’ network
Security analysts say the campaign resembles long-running influence operations attributed to Chinese actors and commonly referred to as “Spamouflage” or “Dragonbridge.”
These networks have been active since at least 2017 and are known for using large numbers of coordinated accounts to promote specific political narratives.
According to experts tracking the activity, the network’s messaging strategy often includes coordinated hashtags, repeated narratives, and the impersonation of credible institutions. In some cases, accounts mimic official campaigns or government messaging to increase perceived legitimacy.
One account cited in the investigation appeared to imitate the U.S. Drug Enforcement Administration’s fentanyl awareness campaign by using similar imagery and naming conventions.
The account simultaneously criticised U.S. drug policy and blamed other countries, including India, for the spread of fentanyl precursor chemicals.
Influence operations beyond Japan
The reports suggest that the influence campaign extends far beyond Japanese domestic politics.
The same network of accounts also targeted Donald Trump, criticising American drug and border policies while shifting attention away from China’s role in global fentanyl supply chains.
According to the research, nearly half of the identified accounts posted messages attacking Trump.
In early February, a coordinated sequence of posts on the fentanyl crisis generated significant online visibility despite the accounts having relatively few followers.
Analysts say such campaigns are designed less to persuade large audiences directly and more to seed narratives that can spread organically through online ecosystems.
Growing security concerns in Asia-Pacific
Cybersecurity experts warn that the scope of Chinese information operations has expanded in recent years, particularly in the Asia-Pacific region.
According to the Google Threat Intelligence Group, Dragonbridge remains the largest pro-China influence network currently under monitoring.
The group reported that China-linked operations have targeted political leadership and public debates in several countries across the region.
Issues such as Japan–Taiwan relations, disputes in the South China Sea involving Vietnam, and political developments in India and the Philippines have all been subjects of coordinated messaging campaigns.
In addition to information operations, investigators have uncovered cyber espionage campaigns linked to Chinese actors affecting dozens of countries.
Security analysts say these activities reflect a broader strategy combining digital influence, cyber intrusions, and online propaganda.
Tokyo’s response and diplomatic tensions
Japanese officials have emphasised that the emerging evidence highlights vulnerabilities in modern information ecosystems. Government representatives say coordinated influence operations could affect electoral integrity and distort public debate.
A spokesperson for Takaichi’s office said authorities are aware of suspicious foreign-linked social media activity connected to Japan’s elections and consider it a serious national security issue.
Chinese officials, however, have rejected the allegations. A spokesperson for the Chinese Embassy in Washington described the research findings as “groundless” and stated that the Chinese government opposes the use of fake accounts or disinformation to manipulate public opinion.
Despite the denial, security analysts say the growing body of evidence from technology companies and independent researchers has intensified international scrutiny of information campaigns linked to Beijing.
Democracy in the age of digital influence
The revelations surrounding the attempted influence campaign highlight the evolving nature of geopolitical competition in the digital era.
As artificial intelligence tools, social media platforms, and coordinated networks of online accounts become central elements of political communication, the potential for external actors to shape narratives across borders has increased significantly.
For Japan, the reports underscore how domestic political debates can become targets within a global information battlefield.
Comment