This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Old Mutual and former finance minister Trevor Manuel have issued warnings regarding deepfake videos circulating on social media. These videos use Manuel's likeness and a cloned voice to promote fraudulent investment schemes. The incident highlights the ongoing use of synthetic media by bad actors to deceive the public into participating in financial scams.
Old Mutual of South Africa reportedly warned of purported deepfake videos impersonating its chairman, former finance minister Trevor Manuel, promoting fake investment schemes on social media. The purported AI-generated videos used his likeness and cloned voice to solicit funds, prompting a public statement from Manuel denying involvement.
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
The deepfake videos generate false information by depicting Trevor Manuel giving unauthorized financial advice.
The incident involves the targeted exploitation of specific victims for financial gain using synthetic media to impersonate a trusted public figure.
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 and subsequent harm were caused by malicious human actors intentionally deploying deepfake technology to defraud social media users.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
Malicious actors utilized deepfake video and voice-cloning technology to impersonate former South African Finance Minister Trevor Manuel in order to promote fraudulent investment schemes. While the incident highlights the growing threat of synthetic media to public trust and financial security, it represents a localized criminal scam rather than a significant national security threat.
The AI system was used as a tool to perform a specific function, namely generating synthetic video and cloning the voice of Trevor Manuel based on direct inputs, without autonomous planning or decision-making.
Limited Risk: The report describes the use of AI-generated content, specifically deepfakes, which falls under Risk Level 3 due to transparency obligations to inform users they are interacting with AI-generated media.
The AI was used to generate deepfake videos and clone the voice of Trevor Manuel to make it appear as though he was giving investment advice.
The report clearly describes the nature of the deepfake scam, the target (Trevor Manuel), and the medium (social media). While specific financial loss figures and victim counts are missing, the classification of the AI risk domain and causal factors is highly straightforward.
None.
Technology
Deepfake Video Generation
AI used to create realistic fake videos of prominent figures.
Artificial Voice Synthesis
AI voice cloning used to mimic the target's voice convincingly.
Unverified AI Trading Software
Claims of automated AI trading software making high-profit decisions.
Data Inputs
Unauthorized Image Use
Scammers used Trevor Manuel's likeness and images without consent.
Fictitious Trading Patterns
Sharing fake trading patterns and signals to lure clients.
Human Factors
Public Trust in Public Figures
Victims trust advice associated with well-known financial figures.
Susceptibility to High Returns
Investors lured by promises of unrealistic or exaggerated returns.
Lack of Verification Habits
Failure of individuals to verify FSP registration before investing.
Process and Methods
Social Media Distribution
Scams distributed easily through social media and Telegram.
Lack of Content Moderation
Social media platforms failed to detect and block deepfake scam videos.
Management
Inadequate Platform Oversight
Social media platforms failed to prevent scam ads and deepfakes.
Regulatory Environment
Unlicensed Financial Advice
Unauthorised entities providing financial services without FSCA approval.
Regulatory Enforcement Gaps
Difficulty in rapidly policing and shutting down online scams.
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.
