1. <Survey on College Students' Awareness of Deepfake Technology Use and Ethics Education>
2. <Improvement Measures for Juvenile Offender Investigation - Focused on the 22nd National Assembly Bills and Changes Related to Deepfake Pornography>

Deepfake, once called a hotbed of crime, is a subcategory of generative AI. This technology, once used only for professional technical demonstrations or limited crimes, has now rapidly popularized to the point where it can be realized on smartphones with just a few touches. In an environment where anyone can easily produce 'deepfake videos,' this article examines together the spread of deepfake crime targeting minors and changes in college students' awareness.


Survey on College Students' Awareness of Deepfake Technology Use and Ethics Education, Park Hee-sook, 2025


Park Hee-sook's paper Survey on College Students' Awareness of Deepfake Technology Use and Ethics Education deals with deepfake crime — the dark side of generative AI — and college students' awareness of it. The author argues for strengthened ethics education and legal punishment through a survey, citing that college students generally perceive deepfake technology negatively (58%) and take its potential dangers seriously (90%). The paper concludes with conclusions that deepfake should mainly be defined as a 'tool of crime' and awareness should be raised. However, closely examining the statistical figures presented by the paper, one can discover that behind the 'negative awareness' interpreted by the author lies a much more complex context.

Deepfake vs. Generative AI: College Students Caught in a Dilemma

The most noteworthy point in this paper's data is the clear gap between 'risk awareness' and 'social impact awareness.' According to [Table 2], for question 7 about the potential dangers of deepfake technology, an overwhelming 90% (99 out of 110 respondents) answered 'very serious' or 'serious,' showing overwhelming concern. In contrast, for question 6 asking about the social influence of technological development, the proportion answering 'negative' dropped significantly to 58% (64 people), with a cautious 'unsure' position reaching as many as 36% (40 people).

The author simplified this to interpret that college students perceive the technology itself very negatively, citing only that positive responses were a mere 5%. However, this 32 percentage point gap (risk 90% vs. negative impact 58%) and high reservation group (36%) is not simple negation. This suggests that college students clearly recognize the criminal abuse possibility of deepfake while experiencing a dilemma between the everyday convenience and positive utility as a creative tool (filters, entertainment, etc.) provided by the 'generative AI' underlying this technology. In other words, college students are not unconditionally rejecting technology but contemplating balance of value judgment between utility and ethics.

Therefore, future ethics education should not stop at simply emphasizing legal punishment for crime prevention or fear appeals as the paper suggests. Already 95% of college students agree with strict legal punishment. Education that is truly needed should focus on developing 'balanced sense' that helps the 36% who answered 'I'm not sure' understand the dual nature of technology and enable them to enjoy the utility of generative AI while setting ethical boundaries themselves.


Improvement Measures for Juvenile Offender Investigation - Focused on the 22nd National Assembly Bills and Changes Related to Deepfake Pornography -

Jang Eung-hyeok · Choi Dae-hyeon, 2025.



The landscape of digital sex crimes has changed. While the past Telegram 'N-th Room' incident was organized crime in a secretive underground world, current deepfake sex crimes have penetrated into classrooms and are spreading like an everyday 'game.' Technological advances have lowered the threshold for crime as long deepfake videos can be created with simple apps, and as a result the age of perpetrators has dropped dramatically. Of the 573 deepfake sex crime suspects apprehended through November 2024, teenagers numbered as many as 463 (80.8%), of which juvenile delinquents aged 10 or older but under 14 — who do not face criminal punishment — accounted for 16.4% (94 people).

The paper Improvement Measures for Juvenile Offender Investigation addresses the gap between the characteristics of digital crime called deepfake and the analog-style juvenile law that cannot keep pace. 

Deepfake Perpetrators Under Age 14

Deepfake pornography was initially a domain of experts targeting famous celebrities on overseas communities like Reddit. However, applications like FakeApp enable anyone with just a smartphone to composite an acquaintance's face into pornographic material. The statistic that 53% of deepfake victims worldwide are Korean shows how deeply this problem has penetrated our society.

The problem is that juvenile perpetrators perceive this not as a serious crime but as simple 'mischief.' According to a Ministry of Education survey, more than half of students (54.5%) answered that they produce deepfakes 'as a prank' or 'out of curiosity (49.3%).' While past juvenile crimes involved physical actions such as violence or theft, current crimes have evolved into a form that destroys a person's character with a few smartphone taps. However, our society's response system remains trapped in the grammar of the past.

Gap in Investigative Authority: Police Cannot Seize a 13-Year-Old's Smartphone

The general public believes that when deepfake crime occurs, police will naturally seize and forensically analyze the perpetrator student's smartphone to find evidence. However, under current law this is nearly impossible.

Under the current Juvenile Act and Criminal Procedure Act system, juvenile delinquents are not subjects of punishment, so police in principle cannot 'investigate' them and can only 'examine' them. The problem is that police have no legal authority to conduct coercive measures — searches and seizures — in this 'examination' process. The core evidence of deepfake crime is data stored in smartphones or PCs. It is not like violence incidents where there are CCTV cameras or witnesses. If a juvenile delinquent refuses saying 'I don't want to show my phone' or hides the device, police have no means to forcibly secure it. Refusal of voluntary submission means that's the end of it.

Ultimately the current legal system fundamentally blocks 'physical evidence securing,' which is the core of digital sex crimes. This goes beyond simply being unable to punish; it prevents the fact of crime from being confirmed at all, leaving victims with no way to resolve their grievances. Police are placed in a legal vacuum where they cannot legally obtain the decisive evidence — the smartphone — even while looking at a boy suspected of the crime.

The 22nd National Assembly's Attempts and Debate: Protection vs. Truth

To solve these contradictions, movements for change are occurring in the 22nd National Assembly. In particular, the juvenile law amendment bill proposed by Representative Kim Seok-ki contains provisions specifying examination authority for juvenile delinquents and, when necessary, granting seizure, search, and verification authority as well as the right to request accompanying warrants. The purpose is to give police the minimum 'weapons' so they can't simply observe crime with their hands tied.

However, voices of concern are also high. Organizations such as the Court Administration Office criticize that giving police strong coercive disposition authority may violate children's human rights and may go against the warrant principle. The counterargument is whether it is really appropriate to treat minors like adult criminals for investigative convenience. Here we encounter a dilemma: should we give up securing evidence to protect children's human rights, or should we allow the intervention of public authority to reveal the substantive truth?

Implications from Japan's 'Separated Response'

The paper focuses on Japan's 2007 juvenile law revision case in solving this difficult problem. Japan also fell into the same dilemma after experiencing shocking juvenile crimes such as the 2003 Nagasaki kidnapping murder case and the 2004 Sasebo elementary school murder case. 

Japan's solution was to thoroughly separate 'physical detention' and 'evidence securing.' The revised Japanese Juvenile Act still did not grant police the authority to arrest or detain juvenile delinquents. Physical freedom was maximally guaranteed for the juvenile's emotional stability. On the other hand, 'search, seizure, and verification' authority for truth-finding was decisively granted to police. That is, it established the principle of 'send the child home, but forcibly secure crime evidence such as weapons or smartphones.'

Also, to prevent secondary damage during the examination process, Japan prepared procedural safeguards such as having psychological specialists 'juvenile guidance officers' participate in examinations and guaranteeing the right to appoint attorneys (assistants). This is evaluated as an attempt to catch two rabbits — 'discovering substantive truth' and 'protecting juveniles' — without leaning entirely toward either unconditional strict punishment or negligence.

Revealing the Truth Is True Protection

Deepfake crime is a heinous crime that destroys victims' souls. If not even the tool of crime (smartphone) can be taken just because the perpetrator is young, that is not tolerance but negligence. The authors conclude that procedural improvements granting legal 'examination authority' and 'search and seizure authority' to police are urgently needed.

Like Japan's example, we need 'targeted legislation' that protects children's bodies while definitively securing crime evidence. Not forcibly extracting confessions but revealing objective facts through digital evidence is the first step to preventing aggrieved victims and also teaching perpetrating juveniles clear responsibility for their actions. It is now time to go beyond 'vague reluctance about giving police authority' and prepare refined legal tools appropriate to the changed crime reality.