Science & Tech · AI · 1 hr ago
South Korea AI regulator received replies to only 29% of risk reports
South Korea’s media and communications regulator sent providers 231 AI risk reports after receiving 401 submissions through a public platform.
Providers replied to just 68 reports, or 29.4%, raising questions about whether voluntary fixes are effective.
Reports included AI giving false legal information and detailed scenarios that could aid crime.
Separately, the regulator received 98 AI-related complaints and reports, most involving Meta’s Instagram and Facebook services.
Rights violations made up 69 of those 98 cases, including privacy breaches and defamation.
The regulator recorded no checks of child-protection measures under a law requiring efforts to prevent children under 14 from receiving inappropriate information.
A lawmaker called for checking what providers do after reports and for a legal basis to verify that chatbot safeguards work.
South Korea’s broadcast and communications regulator received replies to only 68 of 231 AI risk reports it sent to service providers, a response rate of 29.4%.
The reports were among 401 submissions received by a user participation platform by September 30.
The regulator said it cannot force service improvements because its process has no legal binding power.
A lawmaker called for the regulator to verify providers’ follow-up actions and check whether protections for young people work.
The regulator reported no inspections or actions under a law requiring efforts to prevent children under 14 from receiving inappropriate information.
- Who
- South Korea’s Broadcast Media and Communications Commission and AI service providers.
- What
- Providers replied to 68 of 231 risk reports sent to them, a rate of 29.4%.
- When
- The platform received submissions through September 30, 2026.
- Where
- South Korea.
- Why
- The reports concerned risks linked to AI services. The regulator said it cannot compel improvements because its process is not legally binding.
Broadcast Media and Communications Commission
Lee Ju-hee
Legal authority
Broadcast Media and Communications Commission
The commission said its process has no legal binding power, so it cannot force service improvements.
Lee Ju-hee
Lee said the regulator should verify providers’ actions and establish a system for follow-up.
Youth protection
Broadcast Media and Communications Commission
The commission reported no inspection or action record under the relevant law.
Lee Ju-hee
Lee said youth protections should not be left to providers’ voluntary measures and called for a legal basis to check AI chatbots.
The commission’s Online Harm 365 Center received 98 AI service-related cases, including 82 harm reports and 16 submissions.
The AI user participation platform opened.
The platform received 401 submissions, of which 231 were sent to service providers.
Yonhap reported that providers had replied to 68 of the 231 reports.
- Platform submissions
- 401 by September 30, 2026
- Reports sent to providers
- 231
- Provider replies
- 68, or 29.4%
- Most-reported services at Online Harm 365 Center
- Meta’s Instagram and Facebook, 59 cases (60.2%)
- Most-reported harm type
- Rights violations, 69 cases (70.4%)
- Youth protection checks
- No inspection or action record under Article 44-8 of the Information and Communications Network Act
1 source
What people are saying
There is no legal binding force, so we are not forcing service improvements.
Do not stop at forwarding reports; establish a system to check what actions businesses have taken and connect this to follow-up responses.
Do not leave youth protection solely to businesses' voluntary measures; establish an institutional basis to check whether AI chatbot protection measures are actually working.