Singapore media has exposed a network of TikTok accounts using AI to create attractive women who spread distorted information about Singapore and Malaysia, raising fears of undermining public trust and causing social division.
An investigation by CNA found a network of at least 30 TikTok accounts publishing over 550 videos, almost all featuring women created by artificial intelligence (AI) or altered images, with reused voices and scripts to create the illusion that many people are discussing the same issue.
The review revealed that over 98% of the videos were AI-generated or altered, and nearly 90% contained false or misleading content about Singapore and Malaysia, accumulating more than 3 million views in total.
The report stated that these clips have been continuously published from October 2025 to June 2026, mixing factual information with distorted and fabricated content to deceive viewers.
One cited example is a false claim that Singapore's Foreign Minister traveled to plead with China and Indonesia not to open a new shipping route bypassing Singapore's port, an event that never occurred.
After CNA provided information to TikTok for review, two sample accounts were subsequently shut down. TikTok stated they violated policies against deceptive behavior, including manipulative operations, impersonation, spam, and fake reviews.
Experts from Nanyang Technological University (NTU) explained that the goal of such operations is not necessarily to make people believe all the false information, but to have it seen repeatedly until it becomes familiar, causing people to doubt or distrust government information.
Additionally, Malaysia was also targeted, with some clips claiming the country's development is delayed due to policies discriminating against Malaysian Chinese, thereby exploiting sensitive ethnic issues to sow division.
The investigation also found these accounts reused the same scripts, voices, and AI-generated female faces across multiple accounts, posting clips at similar times to create the appearance of multiple sources when they may originate from a single producer.
Experts warned that trying to identify AI-generated content visually may no longer be effective due to rapid deepfake technology advances. Instead, it's important to observe network behaviors, such as repeated messages across multiple accounts or information unsupported by credible news sources.
Academics recommend social media users verify information from multiple sources before sharing, since in the AI era, not only images and voices can be faked, but also non-existent news anchors can be created to subtly influence public opinion and manipulate information.
Click to read news aboutinternational news