Skip to content
Submit incident
AII-1000-72979
AllegedThe record is a lawsuit, complaint or accusation that has not been decided.Supporting sources are pending verification.

South Korean civil servant prosecuted for AI-generated deepfake of colleague

A South Korean civil servant was prosecuted after using generative AI to create a composite image depicting a female colleague as his romantic partner and posting it on KakaoTalk, reportedly causing harm. The prosecution charged him under sexual violence and defamation laws after the victim objected to an initial police investigation that had only applied defamation charges.

November 2025 Β· Seoul Southern District Prosecutors' Office Β· Harmful and abusive content
November 2025
AII-1000-72979
Quick facts
Category
Harmful and abusive content
Country
South Korea
Date
November 2025
Drafted by
Raidu Incident Detection

What happened

A South Korean local civil servant downloaded a photo of a female subordinate from an organizational directory and used generative AI to create a composite image depicting them in a romantic relationship. The image showed the colleague appearing to embrace or rest her hand on the suspect's shoulder.

The suspect posted the manipulated image to his KakaoTalk profile in November 2025. Police initially investigated and referred the case charging only defamation, determining that no explicit sexual exposure or clear sexual conduct was evident.

Reported impact

The victim filed an objection to the police investigation. The district prosecutor's office subsequently reconsidered and added charges under the Sexual Violence Punishment Act for fabrication and distribution of false imagery.

The workplace conducted a workplace grievance review and determined the conduct constituted workplace sexual harassment and harassment. The organization planned further investigation before deciding on disciplinary measures.

Index analysis

The case highlighted ambiguity in legal standards for determining whether synthetic sexual imagery causes sexual shame or humiliation. The Supreme Court had previously ruled in 2021 that sexual shame can be recognized even without exposure depending on context and intent.

Disagreement between police and prosecutors about application of sexual violence laws to AI-generated composite imagery reflected broader concerns that legal frameworks have not kept pace with technology. The 2018 'Hanyang University deepfake case' and subsequent legal amendments were cited as context for ongoing gaps in legal clarity.

Governance controls

The prosecutor stated intent to vigorously maintain prosecution and implement active measures to protect deepfake victims and other sexual crime victims going forward.

Classification

Category
Harmful and abusive content
Industry
Government, security and defence
Country
South Korea
Harm status
Actual harm
Intent
Intentional
Timing
Unknown
Harm types
Human or fundamental rights, Psychological, Reputational
Harmed
Workers
Failure modes
Deepfake image
Consequences
Litigation, Police investigation
AI systems
Generative AI image technology

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident