Translation result.
As policymakers accelerate AI deployment in the name of efficiency and productivity, advocates warn that the rollout must also address harms affecting women, minorities, and care workers — including new forms of gender-based violence.
At the roundtable \”Asking Gender About Social Changes Brought by AI,\” hosted by the Korean Women’s Associations United on May 6 at Seoul Women’s Plaza in Dongjak District, speakers argued that AI is reinforcing existing gendered inequalities. They highlighted problems such as the devaluation of women’s labor, surveillance and isolation in AI-driven care, and the use of deepfakes for sexual exploitation.
Kim Su-a, a professor in Seoul National University’s Department of Communication and Information, opened her presentation, \”AI Use and Social Change: A Gender Perspective,\” by noting that the National AI Strategy Committee’s Korea AI Action Plan emphasizes efficiency and speed. \”Officials often equate AI-driven efficiency with the public interest,\” she said. Kim warned that serious issues are emerging — from the collection of human data to train systems, to fraud and scams, AI in warfare, and deepfakes used for sexual exploitation — and that many of these raise acute gender concerns.
She said early research on AI and gender focused on algorithmic bias and imbalanced datasets. More recent studies show that bias during data collection and evaluation produces errors in assessments of minorities’ and women’s data, undermines worker autonomy, and enables new forms of control. \”Meta, for example, collected real-time data on workers’ keystrokes and hours, which created problems for the quality and ethics of AI training data,\” she said.
Kim added that gender-based harms such as cyberstalking are on the rise and that women’s labor is being devalued. While AI can temporarily increase women’s participation in the labor market, it often perpetuates and widens gaps in income, time allocation, and autonomy. Algorithmic evaluation and rapid AI substitution in care work — a sector with a high concentration of female workers — are intensifying polarization. \”Women’s organizations are calling for structural interventions to protect female-dominated occupations threatened by AI,\” she said.
Kim pointed to the growing use of AI robots in elder care, saying AI check-in call services can track and monitor elderly households and are replacing temporary welfare-center staff who make wellbeing calls. \”We must ask whether this kind of care, measured only by quantity, really counts as good care,\” she said. Kim also warned of a dilemma: while AI chatbots can provide comfort to socially isolated or vulnerable people, they may weaken social ties. Unlike humans, AI bots avoid conflict and offer only consolation — raising the question of whether they are genuinely good conversational partners for care recipients.
Reviewing global regulatory trends, Kim said that accountability for AI firms must prioritize the safety of women, children, and marginalized groups. She noted that Australia has begun assessing chatbots’ effects on children’s mental health and harm and will implement age-based controls for chatbots this year. Australia has also requested transparency and safety reports to determine whether grok bore responsibility for generating child sexual-exploitative images. \”Korea must now consider how to build a culture of responsibility into its AI policy,\” she said.
Lee Min-ju, an activist with the Korean Women’s Association Minwoo, raised concerns about X’s AI chatbot grok. She said that in December 2025 grok added a feature that, when tagged and asked to transform a photo in a particular way, would immediately synthesize an altered image. A French nonprofit analyzed 50,000 grok-tagged posts and 20,000 images over one week that month and found that more than half depicted people in underwear or bikinis, and most subjects were young women. Lee also cited a 2026 case in Seoul in which a district office worker stole a female colleague’s photo, manipulated it to appear intimate with himself, posted it on social media, and sexually harassed her. \”We need to recognize how AI can mediate gender-based violence in new ways,\” she said.
Women’s work replaced by AI: emotional labor has intensified even as the work gets devalued
No Helena, secretary-general of the Korean Women’s Workers Association, said call-center work — a typical early target for AI substitution — illustrates the broader problem. \”AI now handles routine inquiries and repetitive responses, leaving humans to manage complex, emotionally charged consultations,\” she said. That shift does not reduce work; it makes it sharper, increasing emotional labor and skill requirements. Yet because the visible volume of tasks has fallen, employers often devalue the remaining work. Compensation has not kept pace with the increasing difficulty, turning this transitional phase into a more painful form of labor change.
No added that AI does more than replace jobs: during the transition it reshapes work in ways that can be more damaging and reinforces gender stereotypes. Kiosks and AI counselors typically use female voices. \”The technology may look new, but the gender roles embedded within it remain the same,\” she said. Longstanding beliefs that emotional labor and care are women’s responsibilities are being reproduced in technological form, strengthening existing stereotypes.
Most Commented