Full Report
Note to secret agents: ChatGPT is NOT a private diary A ChatGPT user with links to Chinese law enforcement tried to use the AI chatbot to run smear campaigns targeting the Japanese prime minister and other critics of the Chinese Communist Party, according to OpenAI's latest report on malicious uses of its models.…
Analysis Summary
# Threat Actor: Unnamed Individual Linked to Chinese Law Enforcement
## Attribution & Identity
* **Actor Identification:** A specific ChatGPT user with documented links to Chinese law enforcement.
* **Aliases/Groups:** Not explicitly named, but the activities are described as **covert influence operations and transnational repression** likely supported by state-level resources, echoing previous findings often attributed to Chinese state-linked actors (e.g., Meta's attribution of "Spamouflage" to Chinese law enforcement connections).
## Activity Summary
The user attempted to leverage ChatGPT to plan, track, and execute smear campaigns against critics of the Chinese Communist Party (CCP).
* **Recent Campaigns (Mid-October 2025):** An attempt to plan a smear campaign against Sanae Takaichi, Japan's prime minister, after her criticism of China's human rights record in Inner Mongolia.
* **Later Activity (Late October 2025 onwards):** The user documented "cyber special operations" designed to harass and silence both domestic and foreign critics of the CCP. This involved tracking operational progress and generating status reports on these efforts.
* **Specific Incidents:**
* Generating negative comments and coordinating fake complaints using fictional foreign resident accounts against Takaichi.
* Targeting the mental health and families of dissidents.
* Creating fake documentation, such as an obituary and gravestone photos, claiming dissident Jie Lijian had died.
* Attempting to get activist Hui Bo’s X account restricted by submitting thousands of fraudulent violation reports.
* Creating fake claims accusing three dissidents of a sex scandal, which subsequently appeared across multiple online platforms.
## Tactics, Techniques & Procedures
* **Influence Campaign Planning and Execution:** Using AI models (initially ChatGPT, then others) to generate content for political smear campaigns.
* **Impersonation and Deception:** Creating fake email accounts purporting to be foreign residents for official complaints.
* **Coordinated Harassment/Suppression:** Hacking livestreams utilized by dissidents.
* **Abusive Reporting:** Mass reporting of social media accounts (specifically X/Twitter) for phony violations, sometimes supported by fabricated evidence.
* **Psychological Operations (PSYOPS):** Targeting dissidents' mental health.
* **Content Spreading:** Utilizing specific hashtags (e.g., #右翼共生者 - "right-wing symbiont") and posting across multiple platforms (X, Pixiv, Blogspot, Reddit, YouTube, Tumblr, Behance).
* **Operational Documentation:** Using ChatGPT primarily as a private diary or planning tool for "cyber special operations."
## Targeting
* **Sectors:** Political figures/government officials, Chinese dissidents, and general critics of the CCP.
* **Geography:** Japan (specifically the Prime Minister), and global critics of the CCP.
* **Victims:** Sanae Takaichi (Japanese Prime Minister), Jie Lijian (dissident), Hui Bo (activist), and three unnamed dissidents targeted in a fictitious sex scandal.
## Tools & Infrastructure
* **AI Tools:** Initially ChatGPT, later implied use of other LLMs after ChatGPT refused certain requests.
* **Infrastructure:** Mention of launching specific hashtags and generating widespread content across X, Pixiv, Blogspot, Reddit, YouTube, Tumblr, and Behance. The activity suggests **much wider cross-internet activity** than just the prompts seen in ChatGPT.
* **Malware:** None explicitly mentioned.
## Implications
This case demonstrates a **well-resourced and meticulously planned** effort by an actor linked to Chinese law enforcement to use readily available commercial AI tools to conduct **transnational repression**. While the specific smear campaign against the Japanese PM saw minimal traction in the monitored environment, the actor’s broader tactics focus on achieving sophisticated psychological pressure, information warfare (disseminating fake news like obituaries), and systematic digital silencing of critics worldwide. The utilization of LLMs for planning and status reporting highlights a new vector for state-sponsored influence operations.
## Mitigations
* **Enhanced Monitoring:** Increased scrutiny of activity involving specific politically sensitive hashtags and narratives targeting known critics of the CCP.
* **Platform Integrity:** Social media platforms (X, Pixiv, etc.) must enhance detection of coordinated mass abusive reporting campaigns (as seen against Hui Bo).
* **Disinformation Resilience:** Organizations and public figures should be aware that malicious actors are actively using new technologies to generate false content (e.g., fake obituaries, fabricated scandals).
* **LLM Guardrails:** AI providers must continue refining policies to prevent the planning and reporting of influence operations, harassment, and psychological targeting (though the user successfully pivoted to storing plans rather than requesting generation).