She Used AI to Make Nude Photos of Her Boyfriend's Ex. Police Are Now Investigating.

Algorithms, Nude Photos, and a Phone Call: The Law and the Gaps in the Qingdao University AI Case

On September 24, Modern Express called the Discipline Construction and Graduate Training Office at Qingdao University's Art College. A staff member said the college had received a report, and the student involved had graduated in June this year. When asked for details, the answer was: you would have to ask law enforcement.

The same day, a staff member at the Shibei Branch of the Qingdao Public Security Bureau said the victim had reported the matter, the case met the conditions for filing, and an investigation had been opened. The specifics were inconvenient to disclose.

Online posts said a female graduate student surnamed Li at Qingdao University, out of relationship problems, used AI to generate nude photos of her boyfriend's ex-girlfriend, a woman surnamed Wang, and circulated them. She was later criminally detained by police. Police have not fully confirmed these accounts. What is confirmed is this: a report was made, a case was opened, and the graduation date is known. The victim, surnamed Wang, told media she was cooperating with the police investigation and waiting for the result.

The case is still under investigation. How widely the images spread, how many there were, whether there was any profit, and whether other platforms were involved, these key details have not been released. But even based only on what is confirmed, the case has touched several legal boundaries.

The law is there

Using AI to generate nude photos of another person and circulating them may trigger both administrative and criminal liability.

At the administrative level, Article 50 of the Public Security Administration Punishments Law provides that publicly insulting another person or disseminating another person's privacy shall result in detention of not more than five days or a fine of not more than 1,000 yuan. If the circumstances are more serious, detention of not less than five days but not more than ten days, and a fine of not more than 1,000 yuan may also be imposed.

At the criminal level, more than one charge may be involved. The crime of disseminating obscene materials is set out in Article 364 of the Criminal Law. If the circumstances are serious, the punishment is not more than two years of fixed-term imprisonment, criminal detention, or control. The crimes of insult and defamation are set out in Article 246 of the Criminal Law. If the circumstances are serious, the punishment is not more than three years of fixed-term imprisonment, criminal detention, control, or deprivation of political rights.

The Jiangsu High People's Court has published a case. The defendant, a man surnamed Wang, used AI deep synthesis technology to splice the victim's face onto indecent images, fabricated a sexual topic, and posted it on an overseas website. The Zhangjiagang People's Court sentenced him to one year and three months in prison for defamation. The presiding judge said this was a typical case of courts lawfully cracking down on online defamation carried out with artificial intelligence.

There is also a case in Guizhou. After a breakup, a man surnamed Cen paid 250 yuan to have five indecent photos of his ex-girlfriend synthesized and sent them to her relatives and friends, causing her to attempt suicide. The court sentenced him to eight months in prison for insult.

In a Beijing case, a man surnamed Sun made an "AI undressing tutorial" and sold it. The court sentenced him to one year and six months in prison and a fine for illegally using information networks.

There are also cases abroad. A Chinese student in South Korea used deepfake technology to produce more than 1,000 obscene pornographic items and was sentenced by a Seoul first-instance court to one year and six months in prison. A photographer in Taipei used AI deepfake technology to splice the faces of five female musicians onto nude images and published them. The Shilin District Court found that he had seriously infringed on others' privacy.

These cases point in the same direction: whether the image is authentic or fabricated does not affect whether harm is established. Once the victim is recognized, spread, and harassed, the harm has already occurred.

Platform responsibility

The spread of AI-generated content cannot happen without platforms. There are legal answers to whether platforms bear responsibility.

In February 2026, 61 data protection authorities worldwide jointly issued a statement on AI-generated images. The statement noted that AI image and video generation technology has been integrated into widely accessible social media platforms, making non-consensual intimate imagery, defamatory content, and other harmful material possible. It warned that non-consensual intimate imagery may constitute a criminal offense in many jurisdictions, and that developers and users of AI content generation systems must comply with applicable laws.

Domestic regulation is also tightening. In April 2025, the Cyberspace Administration of China launched a special campaign called "Qinglang: Cracking Down on AI Technology Abuse," focusing on issues such as using AI to produce and publish rumors, false information, and pornographic and vulgar content, concentrating on cleaning up related illegal and harmful information and handling and punishing violating accounts and website platforms. In April 2026, the Interim Measures for the Administration of Anthropomorphic AI Interactive Services were published, explicitly prohibiting the use of AI to spread obscene pornographic content, effective July 15, 2026.

A lawyer noted that if social platforms tolerate infringing content such as fake indecent videos synthesized with AI and fail to promptly take necessary measures such as deletion, blocking, and disabling links, they must bear corresponding legal liability.

But platforms facing massive amounts of AI-generated content often cannot keep up with review. It may take only seconds for an image to go from generation to posting. Complaints, review, and deletion, however, must go through multiple steps. In that time gap, the victim has already been seen.

What victims can do

In the past, victims of AI infringement found it very hard to defend their rights. Algorithms, generation records, and backend data are in the hands of platforms and companies, and ordinary people cannot explain them clearly.

In September 2026, the Supreme People's Court issued the Opinions on Lawfully Trying Cases Involving Artificial Intelligence Disputes. This document gives victims more concrete tools on several levels.

On the burden of proof, Article 17 of the Opinions provides that the party controlling documentary evidence and electronic data shall bear the burden of proof. If they refuse to submit it, they shall bear adverse consequences. In determining the validity of evidence of AI-generated content, courts will carefully consider factors such as prompt design, the degree of similarity between the generated content and the work for which rights are claimed, and the consistency of repeated tests.

On prior protection, the Opinions refine the rules for applying personality rights infringement injunctions in AI scenarios. Given that online infringement spreads quickly and consequences cannot be reversed, rights holders are allowed to request a court before litigation to order platforms to delete, block, and take other measures. Si Yanli, deputy director of the Research Office of the Supreme People's Court, said that if a victim has evidence that AI face-swapping was used to create a sexual rumor about them, and failure to stop it promptly would cause irreparable damage to their lawful rights and interests, they may apply to a people's court for a personality rights infringement injunction.

For victims, what they can do is very concrete: first, preserve prompts, generation results, and operation records; send an infringement takedown notice to the platform; call the police; and apply for an injunction when necessary. None of these steps is easy, but each is better than waiting.

AI courses in universities have not caught up

This incident happened at a university, and both parties involved are university students. The problem of AI ethics education in universities has once again been put on the table.

A survey of 1,953 undergraduates at more than 400 universities nationwide showed that AI tool usage had reached 99.18 percent. More than 60 percent of students were confused about the reasonable boundaries of AI use. Less than 30 percent said their school provided systematic AI-related courses.

This means that the vast majority of students, when using AI tools, have never learned in class where the boundaries are. They may know how to enter prompts, but not what they will bear after entering them.

Presidents of several universities talked about AI at opening ceremonies. Presidents of Peking University, Fudan University, and others warned against "cognitive offloading," "false mastery," and "thinking outsourcing" brought by AI abuse. Fudan University President Jin Li said there is both dependence on ready-made answers and a "false mastery" of treating AI-generated content as one's own achievement.

The General Offices of the CPC Central Committee and the State Council's Opinions on Strengthening the Governance of Science and Technology Ethics have made it clear that all science and engineering, computer, and AI-related majors must offer mandatory courses in science and technology ethics and include ethics literacy in student assessment and research evaluation. But implementation still has a gap.

Technology itself has no good or evil. People who use technology need to know where the red line is. A graduate student may be proficient at using AI to generate nude photos of another person, yet may never have learned in class what crime this constitutes and what it means to the victim.

After September 24

On September 24, the phone in the office at Qingdao University's Art College rang. The staff member said the student had graduated in June. A staff member of the Shibei Branch of the Qingdao Public Security Bureau said a case had been opened for investigation, and specifics were inconvenient to disclose. The victim, surnamed Wang, said she was cooperating with the investigation and waiting for the result.

The case is still under investigation. The diploma has already been issued. The case filing receipt is in the victim's hand. The prompts, generation records, and forwarding paths on the phone are still with the police.

After the phone was hung up, the office became quiet again. Someone walked past in the corridor. The bell for the next class had not yet rung.