Deepfake Law in Singapore

Singapore combines a restrictive anti-misinformation framework (POFMA), strong harassment and data protection laws, and a 2023 Online Safety Code that extends moderation obligations to AI-generated content.

Key provisions

POFMA (Protection from Online Falsehoods and Manipulation Act) 2019. Ministers can order correction or takedown of online falsehoods — including AI-generated false-claim content. Applied in numerous instances to deepfake content in Singapore-context political and commercial claims.

Protection from Harassment Act (POHA) 2014. Covers deepfake-enabled harassment, including non-consensual intimate imagery. Updated 2019 to specifically include imagery that appears to depict the complainant.

Personal Data Protection Act (PDPA). Biometric data (voice, facial features) protected. Unauthorized use for deepfake creation implicates PDPA.

Online Safety Code 2023. Published by IMDA (Infocomm Media Development Authority). Large platforms must implement systems to detect and remove harmful content, including AI-generated non-consensual imagery and misinformation.

Penal Code — Cheating, defamation. Standard fraud and defamation provisions apply to deepfake-enabled offenses.

IMDA enforcement

IMDA is the primary media regulator. Enforcement has focused on:

• Directing platforms to remove specific deepfake content.
• Publishing guidance on platform responsibilities.
• Working with Telegram, Meta, TikTok on cross-border enforcement.

Practical implications

For organizations operating in Singapore:

• Platforms: Online Safety Code compliance required for user-to-user services above specified size thresholds.
• AI service providers: PDPA consent requirements; POFMA exposure for content that could be characterized as false.
• Enterprises: moderate-to-strict compliance burden; Singapore enforcement is efficient and predictable.

Related reading

• Japan deepfake laws
• South Korea deepfake laws
• Australia deepfake laws