Kim Soo-hyun AI Voice Review Deemed Inconclusive by Forensic Service

South Korea’s National Forensic Service has reportedly returned an inconclusive finding on whether a disputed recording connected to actor Kim Soo-hyun was manipulated with AI. The Kim Soo-hyun AI Voice issue remains under police review, with reports saying the agency could not determine authenticity because the material submitted for examination was not the original file and contained technical limitations such as noise.¹

The recording was presented by Garo Sero Research Institute as an audio file involving Kim Soo-hyun and the late actress Kim Sae-ron. Kim Soo-hyun’s side has argued that the material was fabricated using AI deep voice technology, while the opposing side has used the file as part of broader allegations surrounding the two actors.²

Forensic Review Finds AI Manipulation Cannot Be Determined

김수현 AI 음성 감정불가 Seoul investigation scene

MBC reported on December 16, 2025, that the Seoul Gangnam Police Station had been notified by the National Forensic Service that it was “unable to determine” whether the audio file disclosed by Garo Sero Research Institute had been manipulated by AI. The report said the agency cited the absence of the original file and noise in the submitted material as factors that limited a technical judgment.¹

The Dong-A Ilbo also reported that police received an inconclusive result from the National Forensic Service in the previous month. The outlet said investigators were looking not only at the disputed audio file but also at other evidence to understand the circumstances of the case.²

The finding does not establish that the recording is authentic. It also does not prove that it was fabricated. In the available reports, the central point is narrower: forensic examiners were not able to make a definitive technical determination on AI manipulation based on the material they reviewed.

That distinction has become important because the dispute is not limited to audio analysis. Kim Soo-hyun’s side filed a complaint against Kim Se-ui of Garo Sero Research Institute on allegations including defamation, arguing that the recording was AI-manipulated. Police have continued reviewing forensic data and related evidence before deciding how to proceed.³

Original File Remains a Key Dispute

A major point in the response from Kim Soo-hyun’s side is whether investigators had access to the original recording. The Korea Economic Daily reported that lawyer Go Sang-rok, representing Kim Soo-hyun, said police had not secured the original file, described as more than one hour long, and that the forensic review was conducted on only a portion played during a press conference.⁴

Financial News, citing News1, reported a similar position from Go, who said the item submitted for analysis was not the original file claimed by Kim Se-ui but only a short sample played at the press conference. In that statement, Go said, “The subject requested for analysis by the National Forensic Service was also merely a sample of a few minutes played by Kim Se-ui at the scene on the day of the press conference.”⁵

MoneyToday reported that Go also said police had received an edited recording of about 50 minutes and requested forensic analysis from the National Forensic Service, while the original file had not been secured.⁶

The dispute over the original file is central because forensic conclusions can depend heavily on the quality, completeness, and provenance of the material examined. The reports provided do not say that the National Forensic Service authenticated the recording. They say the agency could not reach a conclusion on AI manipulation under the available conditions.

Kim Soo-hyun’s legal representative has stressed that an inconclusive forensic result should not be read as clearing the person accused of spreading false information. The Korea Economic Daily quoted Go as saying, “It certainly does not mean that Kim Se-ui’s crime of spreading false information is judged to be without suspicion.”⁴

Police Investigation Continues After Inconclusive Result

The recording at the center of the case was disclosed on May 7, 2025, during a press conference by Kim Se-ui of Garo Sero Research Institute and lawyer Boo Ji-seok, legal representative for Kim Sae-ron’s bereaved family, ChosunBiz reported.³

ChosunBiz also reported that Seoul Metropolitan Police said they were in the process of wrapping up the investigation after receiving forensic results. The police briefing was quoted as saying, “We are finishing the investigation after receiving the results of the forensic examination.”³

MBC reported that police had not ruled out the possibility of manipulation and were expected to decide soon whether to forward Kim Se-ui to prosecutors.¹ The available source material does not include a final police decision, indictment, acquittal, or court judgment.

The current status, based on the cited reports, is therefore unresolved. The National Forensic Service did not determine whether the recording was AI-manipulated. Kim Soo-hyun’s side maintains that the examined material was not the original and that the inconclusive result does not validate the recording. Police are continuing to assess forensic material and other evidence as part of the broader defamation-related investigation.

Kim Soo-hyun AI Voice forensic audio review trend

The case highlights the evidentiary difficulty surrounding disputed voice files in an era of AI-generated audio. For now, the confirmed development is limited but significant: the forensic agency’s result was inconclusive, the original-file issue remains contested, and the police investigation has not been reported as finally resolved.