KUALA LUMPUR – From the prime minister to corporate heavyweights, familiar faces are being digitally hijacked in AI-generated videos promoting bogus investments — signalling a troubling escalation in Malaysia’s cybercrime landscape.
What was once the domain of phone scams and forged documents has evolved into something far more sophisticated.
Artificial intelligence is now being weaponised to fabricate convincing video interviews, cloned audio clips and spoofed news reports, exploiting public trust at scale.
Speaking to Scoop, senior criminologist Datuk P. Sundramoorthy said the surge in deepfake-driven scams represents a fundamental shift in the nature of fraud.
“The recent rise of deepfake videos and fake news site scams in Malaysia reflects an important shift in the nature of fraud and cyber-enabled crime,” he said when contacted.
“From a criminological perspective, this development represents an evolution from simple deception to technologically amplified manipulation, where offenders exploit both human psychology and digital platforms.”
It has been reported that scammers have produced AI-generated videos of Prime Minister Anwar Ibrahim purportedly endorsing high-return investment schemes.
Other individuals depicted in such fabricated videos include former economy minister Datuk Seri Rafizi Ramli, Seputeh MP Teresa Kok, members of senior management at Petronas, and Tan Sri Tony Fernandes of AirAsia.
International personalities have also been misused, including Elon Musk, the chief executive officer of Tesla, Inc.
The deepfake clips are widely circulated on social media platforms and messaging applications before directing viewers to fake registration portals or mobile applications designed to harvest personal information and siphon funds.
Law enforcement has confirmed the threat is real and growing.
It was reported that the Bukit Aman Commercial Crime Investigation Department detected at least five deepfake videos linked to investment scams involving both local and international figures.
The Securities Commission Malaysia has likewise warned that scammers are using AI-generated content to impersonate reputable individuals and companies in order to lend false credibility to fraudulent schemes.
According to Sundramoorthy, the strategy works because it merges two powerful pillars of public trust — authority and media credibility.
“When a familiar face appears in what looks like a legitimate news report, many victims lower their guard,” he said.
He explained that this reflects classic social engineering tactics, where offenders manipulate trust, urgency and perceived legitimacy to influence behaviour.
High-profile individuals are particularly attractive targets because of their visibility and symbolic authority.
“Politicians, corporate leaders and celebrities possess what criminologists describe as ‘borrowed legitimacy’. Their reputation becomes a tool for offenders, even without their knowledge,” he said.
“The more recognisable and respected the figure, the more persuasive the scam becomes.”
The damage, however, extends beyond individuals to institutions.
It was previously reported that The Star issued warnings after discovering fake websites impersonating its brand to promote bogus investments.
More recently, Malaysiakini also became a victim of spoofed content falsely presented as its reporting.
Petronas has cautioned the public against fraudulent schemes misusing its name, underscoring how scammers are hijacking institutional credibility to strengthen their deception.
Such tactics often involve near-identical news portals, spoofed URLs and coordinated networks of fake social media accounts designed to amplify fabricated content, creating the illusion of legitimacy and widespread endorsement.
Unlike conventional scams that require direct interaction, Sundramoorthy said AI-driven fraud can be automated and scaled rapidly.
“A single deepfake can be produced and distributed repeatedly with very little effort,” he said.
“From a routine activity theory perspective, the digital environment increases the availability of suitable targets while reducing effective guardianship, especially where digital literacy remains uneven.”
He added that enforcement agencies face mounting challenges.
“Many of these syndicates operate from overseas using encrypted channels, anonymous domains and cross-border payment systems,” he said.
“Jurisdictional limitations and the rapid pace of technological change complicate investigations. Digital evidence can also be quickly altered, deleted or redistributed across multiple servers.”
Looking ahead, he warned that the threat is evolving.
“AI-powered scams will likely become even more sophisticated,” he said.
“Real-time voice cloning, interactive chatbots and personalised deepfakes could make deception significantly harder to detect.” – February 18, 2026