Purported Widespread Use of AI-Generated Deepfake Videos Impersonate Malaysian Leaders in Investment Scams

Purported AI-generated deepfake videos depicting Malaysia's Prime Minister Anwar Ibrahim and other officials, political leaders, and business leaders have reportedly been used to promote fraudulent investment schemes. A total of 13,956 cases were reported last year, with RM2.11 billion allegedly lost to these scams. Victims were allegedly misled by hyper-realistic fake endorsements circulated via Facebook, WhatsApp, and Telegram. Scammers allegedly adapt quickly, generating new deepfakes when content is removed.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

Scammers are using AI-generated deepfakes and voice clones to impersonate trusted individuals and public figures for illegitimate financial benefit.

Additional risk subdomains

• 3.1 False or misleading information: The AI-generated deepfakes and voice clones present highly convincing false information that misleads victims into believing they are interacting with real people.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The risk is caused by human scammers intentionally utilizing deployed generative AI models to create deepfakes and voice clones for financial fraud.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The reports describe the use of AI-generated content such as deepfakes and cloned voices, which fall under Risk Level 3 due to transparency obligations to ensure users are aware they are interacting with AI-generated media.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Unknown voice cloning technology developers, Unknown deepfake technology developers
• Alleged deployer: Unknown scammers targeting Malaysian public figures, Unknown scammers
• Alleged harmed parties: Tun Tengku Maimun Tuan Mat, Tan Sri Tony Fernandes, Tan Sri Robert Kuok, Malaysian public figures, Malaysian political leaders, Malaysian business leaders, General public of Malaysia, Epistemic integrity, Anwar Ibrahim

Harm severity

Highest direct severity in any category: Severe. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Severe, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Minor
• Privacy: direct Substantial, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Minor, indirect Substantial
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report explicitly describes massive financial losses, stating that RM2.11 billion was lost to deepfake scams last year across 13,956 cases, and RM2.72 million was lost to 454 voice cloning cases.

Directly caused: Victims directly lost money to scammers, with an average financial loss of approximately RM151,189 per deepfake investment scam and RM5,991 per voice cloning scam.

Indirectly caused: N/A

Inferred additional harm: Additional unquantified financial losses are highly likely due to unreported scams and ongoing fraudulent activities that have occurred since the reported statistics were compiled.

Malicious content

Reported: The report explicitly describes the creation and spread of malicious deepfake videos and cloned voice notes.

Directly caused: Scammers directly created and widely circulated hyper-realistic deepfake videos and cloned voice notes on Facebook, WhatsApp, and Telegram to deceive the public.

Indirectly caused: N/A

Inferred additional harm: Thousands of additional deepfake videos and voice notes are likely being generated and circulated continuously as scammers stay ahead of platform takedowns.

Democracy

Reported: The report explicitly describes potential risks to national security and public trust in political leadership.

Directly caused: N/A

Indirectly caused: The manipulation of clips featuring political leaders like Prime Minister Datuk [PERSON_002] could be weaponized to stir panic, manipulate public sentiment, or create political instability, eroding democratic trust.

Inferred additional harm: The long-term erosion of public trust in democratic institutions and official communications is highly likely if deepfakes of political leaders continue to proliferate unchecked.

Privacy

Reported: The report explicitly describes unauthorized voice cloning using public social media videos.

Directly caused: Scammers harvested public audio of individuals from TikTok to clone their voices without their consent.

Indirectly caused: N/A

Inferred additional harm: Widespread unauthorized harvesting of personal biometric data, such as faces and voices from social media profiles, to train and feed generative AI models.

Psychological

Reported: The report explicitly describes emotional exploitation and manipulation of victims.

Directly caused: N/A

Indirectly caused: Scammers use cloned voices of family members to simulate emergency situations, exploiting public emotions and creating a false sense of urgency that causes distress to victims.

Inferred additional harm: Widespread anxiety, distress, and feelings of violation are highly likely among the thousands of victims who fell for these highly personal and manipulative scams.

Epistemic

Reported: The report explicitly describes epistemic harm, noting that deepfakes blur the line between fact and fiction and erode public trust in the media.

Directly caused: The creation of hyper-realistic fake videos of public figures directly misleads the public, making it difficult for viewers to distinguish authentic content from AI-generated fakes.

Indirectly caused: The proliferation of deepfakes erodes overall public trust in media and digital communications, facilitating the spread of false narratives and disinformation.

Inferred additional harm: A broader degradation of the shared information ecosystem in Malaysia, where citizens become increasingly skeptical of all digital media and struggle to maintain a consensus reality.

People affected

• Occurrences reported: 14410
• People reportedly harmed: 14410
• People reportedly exposed: 100000

Potential causes

Technology

• Hyper-realistic Deepfake Videos: AI generates highly convincing videos of trusted public figures to deceive.
• AI Voice Cloning: Scammers clone voices from public clips to impersonate family or friends.
• Rapid AI Content Generation: Scammers quickly generate new fakes to stay ahead of platform takedowns.

Data Inputs

• Publicly Available Media: Scammers harvest short public TikTok videos to clone target voices.
• Lack of Media Verification: Social media feeds lack robust verification of video and audio authenticity.

Human Factors

• Emotional Exploitation: Scams create false urgency, pushing victims to make hasty financial decisions.
• Trust in Public Figures: Victims believe endorsements from trusted politicians and celebrities.
• Lack of Public Awareness: Many users cannot easily distinguish AI-generated fakes from real media.

Process and Methods

• Ineffective Platform Detection: Social media platforms lack strong mechanisms to detect and block deepfakes.
• Slow Content Takedown Process: Scammers outpace manual or slow automated content removal by authorities.
• Weak Platform Verification: Social media platforms often overlook verification of promotional content.

Regulatory Environment

• Loose AI Legislation: Authorities call for tighter legislation against the misuse of AI media.
• Jurisdictional Gaps in Enforcement: Technical response units must refer non-technical fakes to regulators.

Information quality

• Classification confidence: High
• Reason for confidence: The reports from multiple reputable Malaysian news outlets provide consistent statistics on financial losses (RM2.11 billion) and cases (13,956), along with direct quotes from CyberSecurity Malaysia and victims. The role of AI in generating deepfakes and voice clones is explicitly stated, making the classifications highly reliable.
• Ambiguities identified: The specific AI models or platforms used by the scammers are not identified.
• Alternative interpretations: None. The events are clearly identified as AI-enabled financial fraud.

Widespread AI-enabled financial fraud in Malaysia utilizing hyper-realistic deepfakes of public figures and cloned voices of family members. While primarily financially motivated rather than a state-sponsored attack, the scale of economic loss (RM2.11 billion) and the erosion of public trust present substantial challenges to societal stability and economic security.

• Overall national security impact: Substantial
• Response level: Substantial
• Scope: Single nation
• Primary target: Malaysia
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. This represents an ongoing strategic concern and capability deployment rather than an immediate, acute crisis requiring an urgent 72-hour national security response.
• Autonomy: Human-controlled. The AI systems are used as tools directly controlled by human scammers to generate specific deepfake videos and voice clones.
• Novelty: Evolved capability. Represents a significant advancement of existing threat types, scaling up the sophistication and convincing nature of traditional financial fraud.

Impact by dimension

• Physical security: Negligible. No physical infrastructure damage or kinetic threats were reported; the incident is strictly cyber-enabled financial fraud.
• Information security: Minor. While deepfakes of the Prime Minister and other leaders were used, they targeted financial fraud rather than coordinated state-sponsored disinformation or intelligence compromise.
• Sovereignty: Minor. The likeness of the Prime Minister and other officials was exploited for fraud, but there is no evidence of direct compromise to government decision-making or electoral systems.
• Economic security: Substantial. Massive financial losses of RM2.11 billion to deepfake scams represent a substantial economic threat to citizens and financial security, though not yet a systemic collapse of financial systems.
• Societal stability: Substantial. Widespread use of voice clones mimicking family members in emergencies and deepfakes of public figures erodes societal trust and causes significant psychological distress across thousands of victims.