Watchdog asks X to set up minor protection measures for AI chatbot Grok

Seoul, Jan 14 South Korea's media watchdog said on Wednesday it has asked U.S.-based social media platform X to come up with measures to protect minor users from sexual content generated by the artificial intelligence (AI) model Grok.

The Korea Media and Communications Commission (KMCC) said it delivered the request to the operator amid growing concerns over deepfake sexual content that can be generated by AI platforms, reports Yonhap news agency.

"We have asked the operator of X to prevent potential illegal activities on Grok and submit measures to protect teenagers from harmful content, including limiting or managing their access," the KMCC said in a release.

Under the South Korean law, operators of social network platforms, including X, are required to designate an official in charge of minor protection and submit an annual report, the commission said.

The KMCC said the request was made in line with the regulation, noting it has pointed out that creating, circulating or saving sexual deepfake content generated without consent is subject to criminal punishment.

"We intend to proactively support the sound and safe development of new technologies," KMCC Chairperson Kim Jong-cheol said in a release.

"As for side effects and negative impacts, we plan to introduce reasonable regulations and revamp policies to prevent the circulation of illegal information, including sexual abuse content, and require AI service providers to protect minors," Kim said.

Meanwhile, Musk-run X Corp has acknowledged the presence of obscene imagery on its platform, mostly created by its Grok AI, stating that it will comply with Indian laws and remove such content.

The Indian government had directed X to conduct a comprehensive review of Grok's technical and governance frameworks to prevent the generation of unlawful content. It said Grok must enforce strict user policies, including suspension and termination of violators. All offending content should be immediately removed without tampering with evidence, it said.

Disclaimer: This post has been auto-published from an agency feed without any modifications to the text and has not been reviewed by an editor