Wang AI Face-Swap Defamation Conviction (Zhangjiagang, China)

In a case publicised by the Jiangsu High People's Court as one of its 2025 ten typical cases (announced around January 27, 2026), a man surnamed Wang (王某甲) was criminally convicted of defamation for weaponising AI deep-synthesis technology against a female victim.

According to the courts, Wang used 'face-swap' deep-synthesis tools — described as producing images that ordinary members of the public cannot easily tell are fake — to generate indecent images of the victim and to fabricate 'sexual' topics about her. He then repeatedly spread the material through information networks, including into the victim's social and work circles.

The Zhangjiagang Municipal People's Court found that the conduct was serious ('情节严重') and 'seriously harmed social order,' meeting the threshold for criminal defamation as a matter of public prosecution. It convicted Wang of defamation (诽谤罪), treating his voluntary surrender (自首) as a mitigating circumstance. On appeal, the Suzhou Intermediate People's Court rejected the appeal and upheld the judgment.

Commentary accompanying the typical-case listing described it as China's first criminal precedent for defamation committed through AI deep-synthesis, signalling that courts would treat AI 'face-swap' rumour-mongering ('AI换脸造黄谣') as a prosecutable crime rather than a mere civil dispute.

• AI deep-synthesis 'face-swap' tools used to generate realistic indecent images of an identifiable victim from ordinary photographs
• Fabricated 'sexual' rumours attached to the synthetic images to defame the victim
• Material repeatedly distributed across information networks, including into the victim's social and work circles
• Realism of the synthetic imagery made it hard for ordinary viewers to recognise as fake, amplifying the defamatory effect

AI deep-synthesis lowered the barrier to fabricating convincing non-consensual sexual imagery of an identifiable woman, which was then spread online to defame her. Because the images were realistic enough to deceive ordinary viewers, the reputational and psychological damage was compounded.

The documented harm to victims of this conduct includes severe reputational damage, violation of personal dignity through fabricated sexual content, intrusion of the material into the victim's workplace and social relationships, and lasting, hard-to-remedy distress from the network amplification effect. This is third-party harm facilitation: the AI user is the offender and the depicted person is the victim.

• The Zhangjiagang Municipal People's Court (张家港市人民法院, Jiangsu) convicted the defendant, surnamed Wang, of criminal defamation (诽谤罪), finding the conduct '情节严重' (serious) and that it 'seriously harmed social order' — the threshold for public-prosecution rather than purely private defamation. The defendant's voluntary surrender (自首) was treated as a mitigating factor.
• The Suzhou Intermediate People's Court (苏州市中级人民法院) rejected the appeal and upheld the conviction.
• The case was selected as one of the Jiangsu High People's Court 2025 ten typical cases, published around January 27, 2026, and is described as China's first criminal precedent for defamation carried out via AI deep-synthesis technology.
• (Exact custodial term not reliably stated in available reporting.)