Deepfake Elon Musk Videos Have Reportedly Contributed to Billions in Fraud

Scammers used AI to create deepfake videos of Elon Musk promoting fraudulent investment opportunities. Over time, these scams have reportedly led to billions in investor losses. The deepfakes also use voice cloning technology. They have been distributed on social media and YouTube. In particular, they target the elderly, such as 82-year-old Steve Beauchamp, to invest significant sums. Despite efforts by platforms to remove these videos, the scams continue to proliferate.

Risk classification

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

The incident involves malicious actors using AI-generated deepfakes to impersonate trusted individuals and conduct highly targeted financial fraud and investment scams.

Additional risk subdomains

• 3.1 False or misleading information: The AI-generated deepfakes spread highly deceptive and false claims about investment opportunities, leading victims to make decisions based on false beliefs.

Causal factors

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

The incident was caused by human scammers intentionally deploying AI voice-cloning and lip-syncing tools to defraud victims after the models were trained and released.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes the use of AI-generated content such as deepfakes, which are subject to specific transparency and disclosure obligations under the Act.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Unknown deepfake technology developers
• Alleged deployer: Unknown scammers, Unknown deepfake creators
• Alleged harmed parties: Steve Beauchamp, social media users, Retirees seeking investment opportunities, Online investors, Elon Musk, Elderly investors, Cryptocurrency users

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. 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 Minor, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Negligible
• Differential treatment: direct Minor, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report explicitly describes massive financial losses, detailing individual losses of 690,000 dollars, 10,000 dollars, and 36,000 dollars, with industry estimates of billions in losses.

Directly caused: The average financial loss per detailed victim occurrence is approximately 245,333 dollars, based on the three specified victims who lost 690,000 dollars, 10,000 dollars, and 36,000 dollars.

Inferred additional harm: Given that thousands of these deepfake ads are active globally, total financial losses from this specific campaign likely run into tens or hundreds of millions of dollars.

Malicious content

Reported: The report explicitly describes the creation and spread of malicious deepfake videos designed to defraud users.

Directly caused: Scammers created and distributed thousands of malicious deepfake videos of Elon Musk and other celebrities across Facebook, YouTube, and TikTok.

Inferred additional harm: Millions of social media users globally have been exposed to these deceptive and malicious deepfake advertisements.

Differential treatment

Reported: The report explicitly describes how scammers targeted specific vulnerable groups, particularly elderly internet users.

Directly caused: Scammers deliberately targeted older internet users who were familiar with cryptocurrency but unfamiliar with safe online investing practices.

Inferred additional harm: Elderly populations are disproportionately targeted and suffer severe financial and emotional harm from highly realistic AI-enabled scams.

Psychological

Reported: The report explicitly describes psychological distress and self-blame experienced by the victims of the financial fraud.

Indirectly caused: Victims experienced severe distress, embarrassment, and anxiety after losing their life savings, with one victim expressing deep self-blame, writing 'I guess now is the time to call me dumb, stupid, idiot'.

Inferred additional harm: Widespread psychological trauma, anxiety, and depression are highly likely among the thousands of global victims who have been defrauded of their life savings.

Epistemic

Reported: The report explicitly describes epistemic harm through the proliferation of highly realistic deepfakes that erode trust in digital media.

Directly caused: The creation of lifelike deepfake videos of public figures makes it extremely difficult for average users to distinguish authentic footage from synthetic fabrications.

Inferred additional harm: The widespread availability of cheap, convincing deepfake tools threatens to permanently degrade public trust in video evidence and shared reality.

People affected

• Occurrences reported: 1
• People reportedly harmed: 50
• People reportedly exposed: 500000

Potential causes

Management

• Inadequate Ad Screening: Meta and YouTube failed to prevent malicious actors from buying scam ads.
• Cat-and-Mouse Security: Platform security teams struggle to keep pace with shifting scam tactics.

Technology

• Sophisticated AI Tools: Cheap and accessible voice cloning and lip-synching tools released to public.
• Low Cost of Production: Videos cost as little as $10 and take less than 10 minutes to create.
• Imperfect Detection Tools: Current detection websites have high false rates and struggle with compression.

Data Inputs

• Abundant Celebrity Media: Large volume of online interviews of Elon Musk makes deepfaking easy.
• Real Video and Audio Samples: Scammers use genuine clips to train voice clones and lip-sync models.
• Single Image Requirement: Only a single image and video recording are needed to make a deepfake.

Human Factors

• Targeting of Elderly Users: Older users are targeted due to unfamiliarity with deepfake technology.
• Trust in Celebrity Endorsements: Victims believed the realistic Elon Musk videos were genuine endorsements.
• Desire for Quick Wealth: Crypto enthusiasts are drawn to alternative paths for earning fortunes.

Process and Methods

• Paid Social Media Ads: Scammers use paid ads on Facebook and TikTok to amplify their reach.
• Spoofed Live Labels: YouTube live labels are used to make prerecorded deepfakes look real.
• Inconsistent Ad Moderation: Social media platforms struggle to detect and block scam ads in real time.

Regulatory Environment

• Delayed Law Enforcement: Local police departments are overwhelmed and place fraud cases in queues.
• Lack of Platform Liability: Absence of strong regulations holding ad networks liable for scam ads.
• FBI and FTC Warning Limits: Federal agencies warn about deepfakes but lack direct enforcement powers.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide detailed, first-hand accounts from multiple victims, cybersecurity experts, and platform representatives, clearly outlining the mechanics of the scams and the AI technologies used.
• Ambiguities identified: The specific names of the generative AI models used by the scammers are not identified, though their general capabilities are well-described.
• Alternative interpretations: None, as the fraudulent intent and use of deepfake technology are explicitly documented.

Organized criminal groups are leveraging cheap, highly realistic AI voice-cloning and deepfake tools to conduct mass financial fraud globally. While causing significant economic harm to individual citizens and challenging social media content moderation, the incident represents a transnational crime and consumer protection issue rather than a direct threat to state national security.

• Overall national security impact: Minor
• Scope: Multiple nations
• Primary target: No clear primary
• Other affected: Global
• Alleged perpetrator: Organized criminal groups

Threat characteristics

• Imminence: Long-term. Represents an ongoing, persistent criminal capability and strategic law enforcement challenge rather than an immediate national security crisis.
• Autonomy: Human-controlled. AI systems were operated as tools by human scammers to generate synthetic media, with no autonomous decision-making or execution by the AI.
• Novelty: Evolved capability. Represents an evolved capability where existing deepfake and voice cloning technologies are commoditized at extremely low costs to scale fraud.

Impact by dimension

• Physical security: Negligible. No physical systems, critical infrastructure, or kinetic assets were targeted or affected by these financial scams.
• Information security: Minor. While involving mass distribution of deepfakes, the campaign targets individual financial assets rather than state intelligence, elections, or geopolitical stability.
• Sovereignty: Negligible. Core government operations, electoral systems, and state sovereignty remain unaffected by these commercially focused scams.
• Economic security: Minor. Resulted in billions in losses globally and targeted individuals' retirement savings, but does not threaten national-level economic infrastructure or strategic industries.
• Societal stability: Minor. Widespread targeting of vulnerable populations (elderly) causing severe financial and psychological harm, though it does not trigger civil unrest or systemic rights abuses.