The Explosion of AI-Generated Violence Against Women: From Sports Fields to Indonesia's Digital Realm
Baca dalam 60 detik
- Insiden manipulasi video yang menimpa jurnalis ESPN Holly Rowe membuka tabir maraknya konten AI yang menampilkan kekerasan fisik terhadap perempuan di platform digital.
- Algoritma platform dan celah regulasi membuat konten semacam ini sulit ditakedown, sementara riset PBB mencatat 85% perempuan pernah mengalami atau menyaksikan kekerasan online.
- Indonesia menghadapi tantangan serupa dengan UU ITE yang ada, namun belum ada aturan spesifik yang menjangkau deepfake non-seksual, sehingga diperlukan pengawasan platform dan literasi digital yang lebih ketat.

The video manipulation incident involving ESPN field reporter Holly Rowe has sparked a global discourse on the proliferation of AI-generated content depicting physical violence against women. The fake video, showing Rowe being slapped by Texas Longhorns coach Steve Sarkisian, is more than just an edited prank—it's a gateway to a much broader and more alarming phenomenon: thousands of similar clips freely circulating across various social media platforms, ranging from the absurd to the hyper-realistic and disturbing.
Rowe's case originated from a real interaction after the Texas vs. Ohio State game. Sarkisian, unwilling to answer questions directly, left the interview, prompting Rowe's disappointment. The original footage was then modified by an anonymous user to depict a scene of violence that never occurred. The controversy peaked when star Texas quarterback Arch Manning responded to the clip with laughter during a press conference, though he later apologized, and Rowe accepted the apology on September 15.
However, for digital security observers, this incident is merely a surface symptom. Adam Dodge, founder of EndTAB, an organization focused on ending technology abuse, asserts that image-based violence—both explicit and non-explicit—constitutes a major category of online violence against women. "The motives remain the same: to intimidate, objectify, harass, and assert power over the victim," he stated.
While non-sexual violent content may not be as prevalent as deepfake pornography, its traces are easily found. Last year, the tech site 404 Media discovered a YouTube channel containing realistic videos of women being shot. That channel has since been deleted, but dozens of similar videos are still circulating. On TikTok, accounts specifically uploading clips of women being choked—including the character Supergirl—have reached hundreds of thousands of views. There's also a Facebook account with 264,000 followers, ironically named "Dignity Preserved," which regularly posts videos of women being abused with sensational titles.
"What starts online doesn't stop online. Digital violence spills over into real life," asserted Sima Bahous, Executive Director of UN Women.
Alexis Champion from the National Network to End Domestic Violence highlights that such content normalizes violence and reinforces cultural norms that justify oppression. She also points to a unique challenge: non-explicit violent content often does not violate platform community guidelines, making it difficult to remove. "It's hard for the average person to understand why sharing non-explicit images can be harmful," Dodge added.
From a regulatory standpoint, the United States has taken steps through the TAKE IT DOWN Act, signed by President Trump in April 2025. This law mandates platforms to remove non-consensual intimate imagery within 48 hours. However, there is no federal law specifically targeting AI violence videos against women, especially those featuring fictional characters or unidentified individuals. The proposed No Fakes Act, currently under discussion in the Senate, is also limited to the unauthorized misuse of a person's likeness.
For Indonesia, this phenomenon is not a distant threat. With massive internet and social media penetration, deepfake content—both pornographic and violent—has the potential to spread rapidly. While the UU ITE (Electronic Information and Transactions Law) regulates content related to decency and defamation, there is no specific article that covers non-sexual deepfakes. The absence of precise legal frameworks often means that action against AI violence content relies on general interpretation, while global platforms are frequently slow to respond to takedown requests from local authorities.
Furthermore, the low representation of women in AI development—less than 30% according to the WEF—exacerbates the problem. Champion argues that increasing women's voices in technology decision-making spaces would encourage platforms to more aggressively eradicate violent content from within. Without this, existing algorithms tend to be oblivious to local contexts and marginalized victims.
Moving forward, the big question is no longer whether AI can generate violent content, but rather how quickly the digital ecosystem—including Indonesia—can build a fortress of ethics, regulation, and platform oversight. If not, the explosion of such videos will continue to erode women's sense of safety, both in public spaces and on their own phone screens.



