Some 260 fake Jack Ma videos spreading falsehoods about Singapore on YouTube: CNA investigation
The videos – which have already garnered more than 1 million views – use the Alibaba founder’s likeness to deliver a blend of praise and falsehoods about Singapore, to sow seeds of disinformation about the city-state.
As part of the investigation, CNA found that five channels that misappropriated Jack Ma’s image were created within a five-day window and posted on synchronised schedules. Some videos were uploaded by different channels within seconds of one another. (Graphic: CNA/Teha Muhamad)
Read a summary of this article on FAST.
Get bite-sized news via a new
cards interface. Give it a try.
Click here to return to FAST
Tap here to return to FAST
FAST
SINGAPORE: Jack Ma has said little publicly about Singapore in the past decade. But fake, AI-generated versions of the Alibaba co-founder are presenting themselves online as self-professed experts on all things Singapore – from detailed economic facts to urban legends but also outright conspiracies about geopolitical rivalries.
On YouTube, an AI-generated avatar bearing Mr Ma’s face and voice claimed that 10 countries were joining forces to “encircle and suppress” Singapore. Another claimed that Singapore was cutting off other nations’ access to food and water by tightening its control of the Strait of Malacca. CNA found no evidence that Mr Ma ever made these statements, and the videos cited no original speech or interview for such comments attributed to the mainland Chinese entrepreneur.
These videos, which use an authority figure’s likeness to spread falsehoods, can be created at low cost and at scale. An investigation by CNA found 32 YouTube channels that published 300 such videos with avatars of Mr Ma, between September 2025 and June 2026. The videos are in both English and Mandarin.
An AI-assisted review of their content found that 259 – about 86 per cent – contained at least one false or misleading claim in either the videos’ titles or transcripts, or both. These 259 videos, which originated from 25 of the 32 YouTube channels, have chalked up more than 1 million views to date.
Guess Word
Crack the word, one row at a time
Buzzword
Create words using the given letters
Mini Sudoku
Tiny puzzle, mighty brain teaser
Mini Crossword
Small grid, big challenge
Word Search
Spot as many words as you can
These videos are also not obscure content tucked away in the deep reaches of YouTube. In Google searches conducted by CNA in June, the problematic videos appeared in results for English- and Chinese-language queries about “Jack Ma Singapore”, including under the search engine’s Videos tab.
CNA has also found similar content circulating on TikTok, where at least two fake Jack Ma videos have garnered hundreds of thousands of views.
CNA has asked Alibaba if Mr Ma is aware of the videos and whether he plans to take legal action.
INFLUENTIAL TALKING HEAD
Most of these fake Jack Ma videos used one of two recurring formats. The first features a near-static AI avatar of Mr Ma reading from a script, with subtitles below. The second template pairs a still photograph of Mr Ma with an animated audio waveform and a synthetic voice to mimic human narration.
Some channels include disclosures that their content is AI-generated, but not prominently. The videos’ titles and thumbnails still give the strong impression that the real Mr Ma was speaking in the video.
Political scientist Chong Ja Ian of the National University of Singapore (NUS) said Mr Ma’s celebrity status made him an obvious choice.
“The content seems to have a lot to do with money and finance,” he added. “The fact that Ma is a very successful businessperson means that his likeness will draw attention for topics relating to finance, investment, and wealth management.”
The technique of using AI to borrow a public figure’s authority is not new. The likeness of late American investor Charlie Munger was used in the same way in a disinformation campaign against Singapore’s Prime Minister Lawrence Wong.
PRAISE ON A FALSE PLATTER
Most of the 259 videos rarely presented themselves as totally hostile to Singapore. In fact, more than three-quarters had YouTube titles that praised the country.
In other cases, the videos opened with alarmist headlines – one claimed that “3,000 foreign investors” had fled Singapore overnight amid economic collapse – then spent their runtime debunking that premise and extolling Singapore’s competence instead.
In yet other instances, flattery was the wrapping for falsehoods. One video proclaimed that Singapore would be “an absolute safe haven” if World War III broke out, arguing that no rational actor would attack it – and urging overseas Chinese to relocate there.
Associate Professor Chong said this format could lull viewers into treating fake claims as real and unbiased. “These narratives create a false impression and may confuse viewers, especially those less familiar with Singapore,” he said.
He suggested the target audience could be Mandarin speakers across Asia.
“People with means in China are looking for safe locations to store their wealth,” he said. “The positive tone of the videos’ content and their focus on Singapore being a stable, safe location seem most attractive to people who wish to relocate themselves, their families or their money to Singapore.”
A third tactic blended facts with conspiracy. One video started by accurately describing the Deep Tunnel Sewerage System and the Jurong Rock Caverns, before reframing these projects as covert preparations for war amid geopolitical tensions around the world.
Pairing an AI Jack Ma with fact-fiction hybrids echoes a recent CNA investigation, which found over 500 TikTok videos using synthetic female presenters and genuine news hooks to spread baseless claims about Singapore.
In an interview, Professor Anthony Tung from the NUS School of Computing identified two reasons why AI avatars may appear credible to viewers.
First, studies show that people overestimate their own ability to spot synthetic media. Second, with people mostly watching on a small screen and not paying much attention, they make judgments using titles, thumbnails and “a familiar, authoritative face” instead.
In any case, “detection does not equal rejection”, Prof Tung added.
“Many viewers keep watching content they half-suspect is synthetic because it is entertaining, or because it flatters what they already believe – and here, the videos are largely flattering to Singapore, which makes them emotionally rewarding.”
“Repetition then does the damage … Hear the same inflated figure enough times and it starts to feel true, even to those who doubted the source at first.”
MECHANICS OF COORDINATION
CNA’s investigation found that five of the 25 YouTube channels in the 259-video subset showed strong signs of coordinated or automated publishing. Together, these five channels also produced over three-quarters of the content in this investigation.
All five were created within a five-day window in March 2026. Two channels were set up 25 minutes apart, and four posted their first videos within roughly 40 minutes of one another.
On 22 separate days, all five channels then posted within the same 24-hour period, and often on a fixed 15-minute grid such as at 10am, 10.15am and 10.30am.
On at least two occasions, different channels uploaded videos just one second apart.
Their thumbnails also followed a highly consistent formula: Mr Ma’s head and shoulders, two or three lines of Mandarin text and a background image tied to the topic.
Taken together, the signs are consistent with an automated publishing workflow and point to the possibility of a single operator behind it.
“RENTING AUTHORITY”
The 259 videos had drawn almost 1.1 million views, 54,000 subscribers and 31,000 likes and comments combined, as of Jul 1.
Assoc Prof Chong said these numbers sit in the range typical of “pushed content” – where engagement is boosted artificially by bots or coordinated accounts. Even so, he warned that at that volume, should the content ever turn malicious, it could still lead to real reach and amplification.
Additionally, because such videos are typically produced for a fee, the danger lies in how easily their creators can flip their narratives on a whim, experts said.
“These channels can be used to target Singapore, or sow confusion about or in Singapore for the right price – potentially at moments of crisis,” said Assoc Prof Chong.
Four of the 25 channels tracked by CNA are no longer accessible on YouTube. Within the possibly coordinated group of five accounts, one channel has remained active, posting 41 new Singapore-related videos in June alone.
The reach of the videos extends beyond YouTube. A Google search for “Jack Ma Singapore” in June returned fake Ma videos directly beneath the search engine’s AI Overview box. Subsequent checks by CNA found that these fake videos also dominated the top results under the “Videos” tab.
The same search in Chinese surfaces these fake videos under the “Videos” tab too, though they do not appear near the top.
On TikTok, one account posted 12 videos using a similar Jack Ma template. The top clip drew 152,000 views and 5,800 shares. Another TikTok account posted a video that received more than 470,000 views and 13,300 shares – outperforming every individual YouTube video in CNA’s dataset, where the most-viewed video had 56,414 views.
Mr Ma’s perceived credibility may explain some of this attention. But as narrator and presenter, even he is ultimately interchangeable.
One YouTube channel in this investigation has already swapped Mr Ma’s likeness twice: First for the tycoon Li Ka-shing, then for Hong Kong fortune teller Peter So Man Fung – all while keeping its Ma-related name.
Experts said operators often do this to test which faces draw more engagement.
“Swapping the face while keeping everything else constant is the signature of a template-driven content farm,” said Prof Tung, who is also director of the Google-NUS Innovation and Research Centre.
“What is being exploited is authority bias, our tendency to trust the word of authority figures. They are renting authority, not renting Jack Ma.”
Assoc Prof Chong added that the changing faces could also create the impression of multiple content creators, making the network harder to detect – especially when the videos circulate on private messaging apps.
“With enough repetition from enough seemingly different sources, some viewers may come to trust the content – even if they know the images and videos are fabricated.”
Source: CNA/rl(jo)
Sign up for our newsletters
Get the CNA app
Stay updated with notifications for breaking news and our best stories
Get WhatsApp alerts
Join our channel for the top reads for the day on your preferred chat app
Get bite-sized news via a new
cards interface. Give it a try.
Click here to return to FAST
Tap here to return to FAST
FAST
About The Author
