Police Officer in Seoul, South Korea, Alleged to Have Created and Shared Sexually Explicit Deepfakes of Acquaintances

A police officer at Dongdaemun Police Station in Seoul, South Korea, was accused of using an AI app to create roughly 400 sexually explicit deepfake videos from acquaintances' photographs and sharing about 50 through Telegram. He was also charged with possessing about 1,600 child sexual abuse videos. Reporting does not establish whether that material was AI-generated. He was indicted following an undercover police investigation.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

The incident involves the targeted creation of humiliating and sexually explicit deepfake imagery of specific victims (acquaintances) for personal gratification and harassment, which falls under sub-domain 4.3.

Additional risk subdomains

• 1.2 Exposure to toxic content: The perpetrator distributed the sexually explicit deepfakes on Telegram, exposing other users to harmful and non-consensual sexual content.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The harm was directly caused by the intentional actions of a human user (the police officer) who chose to use an AI tool to generate non-consensual sexual deepfakes of his acquaintances.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The AI system was used to generate deepfakes, which falls under the transparency obligations of Risk Level 3 under the EU AI Act.

AI system and alleged parties

• AI system: AI app
• AI purpose: Deepfake Video Generation; Deepfake Image Generation
• Behaviour type: Tool
• Alleged developer: Synthetic video generation technology developers, Synthetic media generation technology developers, Synthetic image generation technology developers, Deepfake technology developers
• Alleged deployer: Synthetic media creators, Law enforcement, Dongdaemun police officer accused of creating sexually explicit deepfakes, Deepfake creators in South Korea, Deepfake creators
• Alleged harmed parties: Victims of non-consensual deepfakes, Privacy, Individuals depicted in deepfake porn, Epistemic integrity, Acquaintances depicted in the Dongdaemun police officer's alleged sexual deepfakes

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Minor, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Minor, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation of 400 deepfake videos depicting acquaintances naked and the distribution of 50 of them on Telegram.

Directly caused: The officer used AI to create 400 sexually explicit deepfake videos of acquaintances and distributed 50 of them on Telegram.

Inferred additional harm: It is likely that these videos were further shared by other users in the Telegram chatrooms, expanding the reach of the toxic content.

Civil rights

Reported: The report explicitly describes violations of sexual violence punishment laws regarding the creation and distribution of false video materials.

Directly caused: The non-consensual creation of sexual deepfakes violated the victims' civil rights and bodily autonomy.

Inferred additional harm: The victims' rights to privacy, dignity, and freedom from sexual harassment were severely violated.

Privacy

Reported: The report explicitly describes privacy violations through the unauthorized use of acquaintances' photos to create deepfakes.

Directly caused: The officer used personal photos of acquaintances without consent to generate 400 sexually explicit deepfakes.

Inferred additional harm: The personal privacy and likeness rights of multiple acquaintances were severely violated.

People affected

• Occurrences reported: 1
• People reportedly harmed: 10
• People reportedly exposed: 50

Potential causes

Management

• Inadequate Monitoring: Lack of internal oversight to detect illegal activities by active officers.

Technology

• AI Deepfake Software: AI tools allowed easy and realistic synthesis of fake nude videos.
• Telegram Platform: Encrypted messaging enabled secure and anonymous sharing of illicit content.

Data Inputs

• Acquaintance Photos: Offender harvested target face images to feed into deepfake AI models.

Human Factors

• Offender Sexual Deviance: Suspect's psychological urge to sexually exploit female and child images.
• Moral Hazard of Officer: Committed digital sex crimes while serving as an active police officer.

Process and Methods

• Flawed Screening Process: Criminal history checks failed to reveal prior sexual offense records.

Regulatory Environment

• Legal Gaps in Disclosure: Suspended indictments are legally omitted from job background checks.

Information quality

• Classification confidence: High
• Reason for confidence: The reports from multiple reputable news agencies provide consistent and detailed factual accounts of the officer's actions, the timeline, the scale of the deepfakes created, and the legal proceedings.
• Ambiguities identified: The exact number of unique acquaintances targeted is not specified, and the specific AI application used is not named.
• Alternative interpretations: None. The facts clearly point to a case of non-consensual deepfake creation and distribution.

A South Korean police officer was arrested for creating and distributing non-consensual sexual deepfakes of acquaintances. While highlighting vulnerabilities in police background checks and representing severe personal privacy violations, the incident remains a localized criminal matter with minor national security implications.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: South Korea
• Alleged perpetrator: Individual

Threat characteristics

• Imminence: Long-term. The perpetrator has been arrested and suspended, resolving the immediate threat, though deepfake abuse remains an ongoing societal challenge.
• Autonomy: Human-controlled. The AI application operated strictly as a tool under direct human control, with no autonomous decision-making or execution.
• Novelty: Established threat. The creation of non-consensual sexually explicit deepfakes is an established threat, though the perpetrator's role as a police officer adds a unique element.

Impact by dimension

• Physical security: Negligible. No physical infrastructure, kinetic targeting, or critical systems were affected by this incident.
• Information security: Negligible. The deepfakes targeted personal acquaintances rather than state leadership, and no classified intelligence or models were compromised.
• Sovereignty: Minor. While the perpetrator was a law enforcement officer, the incident did not compromise core government decision-making, elections, or state sovereignty.
• Economic security: Negligible. There was no impact on strategic industries, financial systems, or national technological competitiveness.
• Societal stability: Minor. The incident represents a severe violation of individual privacy and civil rights for the victims, but lacks the population-scale impact required for a higher national security rating.