Chinese AI Giants Face Regulatory Scrutiny Over Deepfake 'Digital Shaming'
ByteDance's Doubao AI is at the center of a controversy involving the generation of non-consensual pornographic deepfakes through 'fenjue' prompt-injection techniques. Feminist advocacy group Free Nora has called for stricter enforcement as these tools facilitate a large-scale digital shaming campaign against women.
Key Takeaways
- ByteDance's Doubao AI is at the center of a controversy involving the generation of non-consensual pornographic deepfakes through 'fenjue' prompt-injection techniques.
- Feminist advocacy group Free Nora has called for stricter enforcement as these tools facilitate a large-scale digital shaming campaign against women.
Mentioned
Key Intelligence
Key Facts
- 1Doubao records 155 million weekly active users as of late December.
- 2DeepSeek trails as the second-largest competitor with 81.6 million weekly users.
- 3'Fenjue' prompt-injection techniques are being used to bypass AI safety filters.
- 4Free Nora volunteers infiltrated Telegram groups to document the exploitation.
- 5ByteDance has not yet issued a formal response to the allegations.
| Metric | ||
|---|---|---|
| Weekly Active Users | 155 Million | 81.6 Million |
| Safety Vulnerability | High (Fenjue prompts) | Moderate (Reported) |
| Primary Market | China | China/Global |
Who's Affected
Analysis
The intersection of rapid AI deployment and insufficient safety guardrails has reached a boiling point in China, as recent investigations by the feminist collective Free Nora have exposed a systemic exploitation of ByteDance’s Doubao, the country’s leading AI chatbot, to generate non-consensual pornographic deepfakes. This phenomenon, described as 'digital public shaming,' highlights a critical failure in the current regulatory and technical frameworks governing generative AI. As these tools become ubiquitous, the legal and RegTech sectors are facing an urgent mandate to evolve beyond simple content filtering toward robust, adversarial-resistant safety architectures.
The scale of the issue is inextricably linked to the massive adoption of these platforms. According to QuestMobile data, Doubao commands a staggering 155 million weekly active users, nearly double that of its closest competitor, DeepSeek, which holds 81.6 million. This vast user base provides a fertile testing ground for bad actors to experiment with 'fenjue'—a sophisticated prompt-injection system. Named after a secret technique in Chinese fantasy literature, fenjue involves using coded language and specific prompt structures to bypass the safety filters intended to prevent the generation of sexually explicit content. The discovery of these methods within anonymous Telegram groups suggests a coordinated, community-driven effort to undermine platform security.
According to QuestMobile data, Doubao commands a staggering 155 million weekly active users, nearly double that of its closest competitor, DeepSeek, which holds 81.6 million.
From a regulatory perspective, this development exposes the gap between legislative intent and practical enforcement. While China was among the first nations to introduce specific regulations for 'deep synthesis' technology, the Free Nora report suggests that these measures have been largely ineffective in protecting private individuals. Current laws often focus on political stability or large-scale fraud, potentially leaving a vacuum when it comes to gender-based digital violence. For legal professionals, this signals a shift toward more granular liability frameworks where AI developers may be held accountable not just for what their tools are designed to do, but for how easily their safeguards can be dismantled by the public.
What to Watch
The corporate implications for ByteDance and other Chinese tech giants are profound. While ByteDance has remained silent following the South China Morning Post’s inquiries, the reputational risk is mounting. In the global AI race, safety is increasingly viewed as a competitive advantage rather than a mere compliance checkbox. If Doubao continues to be perceived as a primary engine for digital shaming, it could invite heavy-handed intervention from the Cyberspace Administration of China (CAC), which has previously shown little hesitation in throttling platforms that fail to maintain 'social harmony.' This creates a precarious environment for investors and partners who must weigh the growth of these AI models against the legal volatility of the Chinese regulatory landscape.
Looking ahead, the focus for RegTech providers will likely shift toward 'red teaming' and automated forensic tools that can detect fenjue-style prompts in real-time. The battle between AI safety teams and malicious users is becoming an arms race of linguistic ingenuity. For the broader legal community, the Doubao controversy serves as a harbinger of the complex litigation that will inevitably arise from non-consensual deepfakes. As digital identity becomes increasingly fluid, the legal definition of 'personhood' and 'image rights' will require a fundamental overhaul to protect individuals from being 'extracted and sexually degraded' by the very tools meant to enhance human productivity. The case of Doubao is not merely a technical glitch; it is a fundamental challenge to the social contract in the age of generative AI.
Sources
Sources
Based on 3 source articles- Meredith Chen (hk)‘Digital public shaming’: Chinese AI tools under fire for pornographic deepfakesFeb 26, 2026
- Meredith Chen (hk)‘Digital public shaming’: Chinese AI tools under fire for pornographic deepfakesFeb 26, 2026
- Meredith Chen (cn)‘Digital public shaming’: Chinese AI tools under fire for pornographic deepfakesFeb 26, 2026
Cite This Page
"Chinese AI Giants Face Regulatory Scrutiny Over Deepfake 'Digital Shaming'." Legal & RegTech Intelligence Brief, February 27, 2026. https://getlegalbrief.com/story/chinese-ai-deepfake-regulation-crisis
How we covered this story
Every story in our legal coverage is assembled from multiple primary sources, cross-referenced for factual consistency, and scored along three independent dimensions: sentiment, operational impact, and source-cluster confidence. Single-source rumors and unverifiable claims do not pass our editorial gate. When a story shows "Verified by N sources" with N≥2, the development is independently corroborated; when N=1, we mark it explicitly so readers can weigh the signal accordingly.
Impact scoring uses a 1-10 scale weighted toward regulatory, financial, and operational consequence rather than coverage volume. A topic that runs in every outlet but moves no real decisions ranks lower than a niche regulatory filing that reshapes how operators in the legal space have to behave. Read our full methodology for the scoring rubric, our glossary for term definitions, and our trends index for the longitudinal view across the beat.
Sources are only linked to a story once they clear our classification pipeline at a minimum 35 percent relevance threshold. According to that methodology, reviewed July 2026, this follows multi-source corroboration standards recommended by journalism research bodies such as the Reuters Institute for the Study of Journalism.
See something wrong in this story — a wrong fact, a broken source link, a misattributed entity? Report a data issue.
| Signal on this page | What it tells you |
|---|---|
| Verified by N sources | Independent corroboration count. N≥2 is our confidence floor; N=1 is marked explicitly. |
| Impact score (1-10) | Regulatory + financial + operational weight. 8+ signals an experienced-operator action item. |
| Sentiment | Five-tier classification trained on labeled legal-specific corpora. |
| Timeline | Where applicable, the related-events sequence that contextualizes today's development. |