SEOUL (AFP) – South Korea’s national police said yesterday that they planned to spend nearly USD7 million over the next three years on technology to battle deepfakes, voice cloning and other forms of digital fraud.
In recent weeks, authorities have uncovered a sprawling network of Telegram chatrooms, often set up within schools and universities, in which users shared artificial intelligence (AI)-generated deepfake sexual content depicting female students and staff.
The revelations have prompted public outrage, with the president vowing stern action.
Seoul’s National Police Agency said yesterday that it had allocated KRW2.7 billion (USD2 million) a year until 2027 for developing deep-learning technology to detect digitally-fabricated content such as deepfakes and voice cloning.
In addition, they would also spend hundreds of thousands of dollars to upgrade their current software for monitoring deepfake and other AI-generated videos.
Super-wired South Korea has long battled sexual cyber-violence, but cases are up 11-fold this year from 2018, official figures show.
Prosecution rates, however, remain dismal. From 2021 to July this year, 793 deepfake crimes were reported, but only 16 people were arrested and prosecuted, according to police data obtained by a lawmaker.
South Korean President Yoon Suk Yeol said last week that the “issue of deepfakes has sparked widespread concern, particularly among women, as it represents a serious crime that undermines social harmony”.
“I urged the relevant ministries to take decisive action,” Yoon added.
In late August, 84 women’s organisations released a joint statement saying the root cause of the deepfake crisis was “structural gender discrimination and the solution is gender equality”.