AI-Generated Deepfakes Deployed in Coordinated Disinformation Campaign Targeting Singapore on TikTok
Key Takeaways
- ▸AI-generated personas have enabled factory-scale disinformation: 550+ coordinated videos reaching 3M+ viewers with recycled synthetic faces and voices
- ▸The campaign strategically blended real news as narrative hooks (ExxonMobil closure, Hainan port launch) with fabricated geopolitical claims to create plausible-sounding false narratives
- ▸Nearly all videos (98%) employed AI-generated or manipulated female personas, demonstrating how generative AI has commoditized the creation of synthetic news presenters
Summary
A Channel NewsAsia (CNA) investigation has exposed a systematic disinformation operation using AI-generated female personas to spread false narratives about Singapore and Malaysia on TikTok. Between October 2025 and June 2026, approximately 30 coordinated accounts produced over 550 videos reaching more than 3 million viewers. The operation employed a factory-like system where 98% of videos featured AI-generated or manipulated female personas paired with recycled text-to-speech voices and reused scripts. Nearly 90% of videos contained false or misleading claims strategically designed to portray Singapore's economic and geopolitical decline—including a fabricated story that Foreign Minister Vivian Balakrishnan begged China and Indonesia to preserve Singapore's port dominance. The campaign weaponized real events (ExxonMobil plant closure, China's Hainan Free Trade Port expansion) as hooks for outright fabrications, despite Singapore's port handling a record 44.66 million containers in 2025. Following CNA's investigation, TikTok removed at least two identified accounts. Parallel investigations suggest similar coordinated AI-generated disinformation campaigns operate across YouTube and other platforms.
- Multi-platform coordination detected: similar campaigns identified on YouTube and other platforms, suggesting organized infrastructure for AI-enabled influence operations
Editorial Opinion
This investigation reveals a troubling new frontier in AI-enabled disinformation: the industrial-scale production of synthetic personas to launder false narratives. The operation's sophisticated use of real-world events as scaffolding for outright fabrications—combined with the commodity cost of AI-generated video and audio—demonstrates how generative AI has dramatically lowered barriers to coordinated influence campaigns. As these tools become more accessible and harder to detect, the speed of platform response (TikTok's removal) matters less than the underlying arms race: detection and attribution must keep pace with AI synthesis capabilities, or coordinated disinformation will outpace remediation.
