China Weaponises AI to Hunt Religious Dissidents Worldwide?

 

Anthropic’s September 10 report says China-linked operators used Claude to gather intelligence on Falun Gong practitioners, religious leaders and overseas dissidents, producing thousands of dossiers from multilingual social-media data for potential state-security use at scale.

In a stark revelation of Beijing’s expanding digital repression, Chinese state-aligned actors have been caught using Anthropic’s Claude AI system to conduct large-scale intelligence gathering on religious communities and overseas dissidents. The targets include Falun Gong practitioners, Tibetan Buddhists, Taiwanese Presbyterian leaders, Asian Catholic cardinals, and other groups long viewed as threats by the Chinese Communist Party (CCP).

Anthropic’s September 10 threat intelligence report details how these operators treated Claude as an “intelligence collection desk.” One account identified itself as belonging to a Chinese state information security officer. The company assessed the activity as a China-based intelligence operation closely aligned with the priorities of China’s religious affairs apparatus, the Ministry of State Security, and the United Front Work Department, the CCP body tasked with co-opting diaspora communities and neutralising critics abroad.

From a single machine, one operator generated 2,475 finished intelligence products, dossiers, reports, and daily digests in just 30 days. Claude ingested material in multiple languages from both Chinese platforms (WeChat, RedNote, Douyin, Weibo) and Western ones (LinkedIn, Instagram, Threads, X, Facebook), then translated, summarised, and produced polished Chinese-language outputs. The operators explicitly instructed the AI to adopt “China’s standpoint” and employ official CCP terminology such as “illegal separatists.”

The surveillance was systematic and intrusive. Targets included senior Catholic cardinals across Asia, Presbyterian Church leadership in Taiwan, Tibetan exiled leaders, Falun Gong practitioners. Operators sought sensitive personal details like birthdays, immigration dates and attempted to map religious venues using floor plans, facade images, and structural diagrams.

They specifically requested information on China-related activities, “negative information,” and personal or organisational vulnerabilities that could be exploited for leverage. The daily reporting cycle appeared designed for internal state security consumption.

Falun Gong, a spiritual discipline based on truthfulness, compassion, and forbearance, has faced relentless persecution in China since the CCP launched its nationwide crackdown in July 1999. Practitioners have endured arbitrary detention, forced labour, torture, and organ harvesting.

The Anthropic report notes that one surveilled individual is an instructor at Fei Tian College’s Northern Campus in Ontario, Canada an institution linked to Shen Yun Performing Arts, the classical Chinese dance company founded by Falun Gong practitioners that presents traditional Chinese culture free of Communist ideology and highlights human rights abuses under CCP rule.

This AI-assisted operation fits into a broader pattern of transnational repression. U.S. - based Falun Gong practitioners and Shen Yun affiliates have reported hundreds of incidents since 2020, including physical assaults, property damage, threats, and harassment of relatives still in China. In dozens of cases, surveillance was involved. Anthropic also identified accounts linked to municipal public security and state security organs that sought “pre-operational venue intelligence” on overseas protests, Uyghur cultural events, and gatherings such as the Oslo Freedom Forum.

Anthropic has banned the accounts and strengthened its detection systems. The episode underscores a growing risk: sophisticated state actors are already testing and exploiting advanced AI tools to multiply the reach of traditional intelligence work. What once required teams of analysts can now be scaled rapidly by a single operator with access to a capable language model.

For democratic societies and free religious communities, the message is clear. The Chinese Communist Party continues to export its domestic repression through technology, targeting not only its own citizens but also diaspora groups and faith communities that refuse to submit to Party control.

Vigilance against such AI-enabled transnational operations is no longer optional, it is a necessity for national security and the defence of fundamental freedoms.