Are we living in a deepfake ‘Truman Show’? — Lim Shining and Mohd Istajib Mokhtar

On August 12, an anonymous Telegram message sent to Heejin warned her of leaked pictures and personal information. Included in the message were images of her high school photo and a deepfake featuring her face digitally swapped onto a stranger's naked body. Unfortunately, Heejin is not alone in this crisis. A BBC News investigation revealed a deepfake pornography epidemic in South Korea, primarily targeting young women and children across more than 500 schools and universities.

Deepfakes are synthetic media created with artificial intelligence (AI) to replace one person's likeness with another's. This technology is increasingly being weaponized to produce highly realistic, sexually explicit content without consent. The South Korean "Nth Room" case still echoes with trauma, but the issue is widespread globally.

In China, a Telegram group called the "MaskPark Treehole Forum" had over 100,000 male members operating as a hidden marketplace for trading AI-generated deepfake pornography and other non-consensual intimate imagery.

In Malaysia, the crisis hits closer to home. A 16-year-old male student in Kulai, Johor, was arrested for generating explicit deepfake images of over 40 female schoolmates and alumni. He sold these synthetic images in a Telegram group for just RM2 each. The school authorities had blamed the victims, claiming "girls shouldn't post photos online", sparking outrage over the institutional failure.

This weaponization of AI affects everyone, targeting women in positions of power. In 2022, Northern Irish politician Cara Hunter was targeted with an explicit deepfake video distributed on encrypted apps just weeks before her election.

The creation and dissemination of non-consensual deepfake pornography are ethically and legally problematic. Researchers refer to this as "digital misogyny" - an online culture that views women as mere raw materials to be manipulated and consumed. Legally, loopholes abound. The mere existence of deepfakes allows guilty parties to dismiss genuine incriminating footage as fake, undermining the foundational legal concept of evidence.

Additionally, a copyright loophole exists due to the Thaler v. Perlmutter ruling in the US, which states that AI-generated works without significant human authorship are not copyrightable.

Malaysia is grappling with outdated laws to combat 21st-century crimes. The Penal Code's Section 292 on obscenity and Section 509 on insulting modesty are strictly based on real physical bodies. Victims with their faces superimposed onto synthetic bodies often find no clear legal recourse. The Communications and Multimedia Act (CMA) 1998 can only act as a reactive measure after reputational harm has already occurred.

The path forward involves standardizing AI legislation globally, making unauthorized digital replicas of a person's appearance or voice a high-tier cybercrime, and updating Malaysia's penal codes to explicitly define and criminalize the creation, possession, and distribution of AI-generated deepfakes.