Korean women remove pictures, videos from social media amid deepfake porn crisis
Growing misuse of AI-generated “deepfake” porn is prompting Korean women to wipe their images from social media and has led South Korea’s parliament to criminalise even possession and viewing of such material. Commenters debate whether stronger laws can meaningfully deter abuse when image and video generation is cheap, anonymous and hard to police, or whether society will simply adapt to a world where visual media is inherently untrustworthy. The thread also raises broader concerns about harassment, gendered impacts, free speech, and how deepfakes may erode the evidentiary value of photos and videos in everything from personal disputes to politics.
Scope and context (South Korea, surveillance, gender)
- Commenters link the deepfake crisis to South Korea’s existing problems with hidden cameras in hotels/toilets and pervasive state and CCTV surveillance.
- Some worry authorities could fabricate incriminating footage of anyone.
- One thread ties the issue to South Korea’s skewed male–female ratio and tense gender relations.
How serious are deepfakes?
- Some see deepfakes as the most acute near-term AI risk, due to ease of use and scale.
- Others argue they’re inevitable and society will adapt; saturation will make people distrust all images/videos.
- A minority argues they’re “not a concern at all” and might even weaken revenge porn by undermining evidentiary value. Critics counter this ignores social stigma and irrational reactions.
Harassment, victims, and gendered impact
- Many emphasize the psychological harm of being targeted, regardless of whether others know it’s fake.
- Examples include stalking-style campaigns against individuals, fear of escalation, and job loss from sexualized images (real or fake).
- Several note that obvious harassment vectors weren’t considered in AI development, reflecting underrepresentation of women in tech.
- Disagreement over framing harms via “imagine if it were your daughter”; some see it as effective persuasion, others as paternalistic.
Barrier to entry vs Photoshop
- Strong consensus that AI tools drastically lower skill/time required compared with Photoshop, enabling mass production and making average users potential abusers.
- A minority contends that anyone competent enough to use modern AI tools could already do convincing Photoshop, so the bar is similar; others strongly dispute this.
Law, regulation, and free speech
- New Korean law penalizing possession/viewing of deepfake sexual material raises concerns about criminalizing mere file storage.
- Some argue existing harassment/revenge-porn laws should apply; others say enforcement is already very hard, especially with anonymity and cross-border attacks.
- Debate over whether restrictions would meaningfully infringe freedom of speech, and whether that matters here.
Trust in media and technical fixes
- Widespread concern that photos and videos will lose evidentiary value, affecting everything from politics to criminal justice.
- Proposed mitigations: watermarking, cryptographic signing from trusted devices; critics note spoofing, the “analog hole,” and that trust is ultimately institutional, not purely technical.