Hong Kong is considering whether creating a sexually explicit deepfake for private use, without publishing it, should itself become a criminal offence. Secretary for Justice Paul T K Lam said officials would have to weigh whether the act harms another person and how far the law should extend when no third party has been directly affected.
The question forms part of a wider plan to regulate specific AI risks rather than introduce one comprehensive set of controls. The latest Policy Address identified protection for underage users and AI-related crime among the areas for attention. Lam said the aim was to protect people without obstructing useful AI development.
Deepfake legislation would also have to settle what counts as an image under existing law. Officials are examining whether an AI-generated image falls within legal references to a photograph, and how to define covered material clearly enough to draw a line around unlawful conduct. Lam said rules drafted too broadly could be unclear, while rules drafted too narrowly could become outdated as the technology changes.
The Law Reform Commission of Hong Kong is reviewing existing law on AI-related crime. Lam said the commission was considering whether AI-related cybercrime should be separated from the wider review of computer crime and given priority. He personally hoped a preliminary proposal could be put out for public consultation as early as the end of 2026.
Officials are preparing a cross-departmental task force to examine minors’ use of social media and AI. They will consider whether regulation should apply to children, parents or platform providers, a choice that Lam said would affect how workable any measure was.
Hong Kong has not chosen a uniform social-media age threshold. Officials will study approaches elsewhere, including Australia’s reported restrictions for people under 16, alongside possible platform measures such as content filters, disabling autoplay and limiting infinite scrolling.