This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Singapore Prime Minister Lawrence Wong has issued a public warning regarding the proliferation of deepfake videos and audio clones of himself being used to facilitate financial scams. These manipulated assets, which impersonate the Prime Minister to promote cryptocurrency and other fraudulent schemes, have been identified on social media platforms. The government has urged citizens to remain vigilant and report such content via official channels like ScamShield, noting that this is part of a broader, ongoing trend of AI-enabled impersonation targeting Singaporean leadership.
Singapore Prime Minister Lawrence Wong issued a warning about AI-generated deepfake videos and voice clones falsely portraying him promoting cryptocurrency scams, money-making schemes, and PR services. The manipulated content, seen on social media, reportedly uses public footage and AI voice cloning to deceive victims. Wong urged the public to avoid engaging, report scams via ScamShield, and stay vigilant.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Also classified as
Deepfakes of Senior Minister Lee Hsien Loong commenting on international relations were created to mimic government positions, posing risks to national interests and public opinion.
The deepfakes and voice clones were created by scammers to impersonate public figures for financial gain through fraudulent cryptocurrency and investment schemes.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The risk is caused by human malicious actors intentionally using post-deployment AI generation tools to create deepfakes for financial fraud.
Added categories peak at 3: Substantial (Epistemic Harm, Inferred)
The proliferation of AI-generated deepfakes and voice clones impersonating Singaporean leaders for financial scams and fabricated commentary on international relations poses a substantial threat to official state communication and public trust. While primarily a criminal fraud issue, the potential for political manipulation elevates its national security significance.
The AI systems (deepfake generators and voice cloners) were used as tools by malicious actors to perform specific, well-defined functions of generating synthetic video and audio of Singaporean leaders based on direct inputs.
Risk Level 3. Limited Risk: The report describes 'AI-generated content such as deepfakes' which pose a moderate risk and require specific transparency obligations to ensure users are informed.
The AI systems were used to generate deepfake videos and cloned voices of Singaporean leaders to impersonate them in scam advertisements.
The reports clearly describe the nature of the incident, identifying the targets (Lawrence Wong and Lee Hsien Loong), the platforms used (Facebook, X), and the intent (financial scams and political impersonation). The role of AI (deepfakes and voice cloning) is explicitly stated.
None. The reports consistently describe these as malicious deepfake scams.
Technology
AI Voice Cloning
Scammers used AI to clone the Prime Minister's voice convincingly.
Advanced Video Manipulation
AI algorithms altered facial movements and expressions in real videos.
Data Inputs
Exploitation of Public Footage
Real interviews of leaders were harvested to train and create deepfakes.
Human Factors
Trust in Authority Figures
Citizens trusted the manipulated videos showing their national leaders.
Low Detection Capability
Average users struggle to spot subtle anomalies like atypical blinking.
Process and Methods
Algorithmic Feed Distribution
Social media algorithms promoted deepfake ads on user feeds.
Reactive Incident Response
Mitigation relied heavily on user reports and warnings after exposure.
Management
Inadequate Platform Screening
Social platforms failed to block deceptive ads using deepfaked public figures.
Regulatory Environment
Cross-Border Jurisdictional Gaps
Online scams often originate remotely, hindering local police enforcement.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
