AI Virtual Companion Apps in China Expose Minors to Sexual and Violent Content
Multiple AI virtual companion apps in China reportedly failed to prevent minors from accessing sexual and violent content, despite regulatory guidance. Testers found that apps including EchoMe and 筑梦岛 allowed users to bypass teen protections, select extreme storylines, and access sexually explicit material after disclosing their age.
What happened
In April 2026, China's National Internet Information Office and four other departments published the Interim Measures for Managing AI Personified Interactive Services, prohibiting the provision of content that is obscene, pornographic, violent, or likely to induce extreme emotions in minors. However, multiple AI virtual companion apps including EchoMe, 筑梦岛 (Zhuimeng Dao), and 喵呜小手机 (Miaowu Xiaoshouji) were found to violate these regulations.
The apps displayed system preset characters with explicit imagery, allowed users to select extreme storylines including violence, infidelity, and divorce scenarios, and contained sexually suggestive language. A 13-year-old tester was able to bypass the teen mode with a single click, required no age verification during registration, and received responses encouraging romantic involvement after disclosing their age. A third-party market developed on secondhand trading platforms offering character customization services, with pornographic character designs available for 13.14 yuan.
Reported impact
Minors were exposed to sexual and violent content through apps marketed as emotional companions. The apps employed engagement tactics including emotional manipulation and paywall systems requiring charges for features such as voice playback, storyline unlocking, and accessing character profiles.
The apps featured characters with sexually suggestive labels and descriptions, including tags such as 'empty', 'depressed', and 'domination/submissive', alongside explicit imagery of exposed bodies and vulgar language describing intimate acts.
Index analysis
The apps failed to implement effective age verification or teen protection mechanisms, instead allowing one-click bypass of teen mode options. Content moderation did not prevent exposure to extreme storylines and sexually explicit material to identified minors.
The development of an external market for character customization, including access through API integration, expanded the scope of potentially harmful content beyond platform moderation.
Governance controls
In April 2026, the National Internet Information Office and four other Chinese departments published the Interim Measures for Managing AI Personified Interactive Services, scheduled to take effect on July 15, 2026. The measures explicitly prohibit offering virtual companion services to minors and require effective measures to identify minors and inform users they are interacting with AI rather than humans.
Legal experts stated that platforms must implement content review and data security responsibilities, and that third-party customization and API integration must comply with regulatory requirements. They noted that inducing minors to make charges could result in contract invalidity and administrative penalties.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.