This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A coordinated financial fraud campaign in Israel utilized AI-generated deepfake videos of prominent public figures to promote a stock pump-and-dump scheme. The scam, which targeted investors through social media and messaging platforms, resulted in significant financial losses for victims after the stock value collapsed. The incident highlights the exploitation of digital advertising tools and the challenges platforms face in moderating AI-driven impersonation content.
A coordinated fraud campaign in Israel allegedly used AI‑generated deepfake impersonations of public figures including Amir Yaron, Benjamin Netanyahu, Eyal Golan, Noa Kirel, Gal Gadot, Elon Musk, and Mark Zuckerberg to promote NASDAQ‑listed Ostin Technology Group Co. Ltd. (OST) in a pump‑and‑dump scheme. Paid social media ads, WhatsApp groups, and fake investment platforms reportedly lured investors before the stock collapsed in early July 2025, causing major reported losses.
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 generated and spread false information regarding financial endorsements to mislead investors.
The scammers used AI-generated deepfakes to impersonate trusted individuals for financial gain through a fraudulent stock scheme.
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 incident was orchestrated by human fraudsters who intentionally deployed AI deepfakes and ad tools to execute a financial scam.
Added categories peak at 3: Substantial (Epistemic Harm, Inferred)
A coordinated financial fraud campaign in Israel utilized sophisticated AI-generated deepfakes of prominent figures, including Prime Minister Benjamin Netanyahu and central bank officials, to execute a stock pump-and-dump scheme. While causing significant financial losses for individual victims and highlighting vulnerabilities in social media ad moderation, the incident has minor national security implications, representing an evolved cyber-fraud capability rather than a state-sponsored threat.
The AI systems (deepfake generators) were used as tools to produce synthetic videos based on direct inputs, without exhibiting independent goal-directed planning or autonomous decision-making.
Risk Level 3: Limited Risk. The report describes the use of AI-generated content, specifically deepfakes, which falls under the transparency obligations of the EU AI Act for synthetic media.
The scammers primarily used AI to generate highly realistic deepfake videos of public figures to promote the fraudulent stock.
The report provides clear details about the nature of the scam, the platforms used, the specific public figures impersonated, and the financial impact on victims. The role of AI (deepfakes) is explicitly stated.
None
Technology
Weaponized Deepfakes
AI-generated videos of public figures used to build trust and mislead.
Dynamic Ad Delivery
Ad tech showing scam content only after user clicks, bypassing filters.
Fake Trading Platforms
Slick external websites mimicking real dashboards and market graphs.
Data Inputs
Targeted Demographic Data
Exploiting platform ad tools to target specific demographics with precision.
Opaque Advertiser Identities
Lack of verification for advertiser profiles and identity data.
Real Stock Graph Data
Using real rising stock graphs to provide a false illusion of proof.
Human Factors
Trust in Impersonated Figures
Victims believed deepfakes of trusted finance and public figures.
Exclusivity and FOMO
Scammers used WhatsApp groups to build trust and a sense of exclusivity.
Lack of Deepfake Awareness
Victims unable to distinguish AI-generated videos from real footage.
Process and Methods
Piecemeal Content Moderation
Meta removed individual posts instead of taking down whole scam networks.
Poor Cross-Ad Tracking
Inability to track and block cloned scam ads after a takedown.
Easy External Redirection
Allowing ads to seamlessly redirect users to unverified external sites.
Management
Prioritizing Ad Revenue
Platforms profiting from sponsored scam ads, reducing incentive to act.
No Proactive Infrastructure Block
Failure to invest in systemic tools to block coordinated scam networks.
Inadequate Risk Assessment
Failure to recognize and mitigate early signs of malicious AI content.
Regulatory Environment
Weak Oversight and Enforcement
Local regulatory gaps left victims vulnerable to international schemes.
Lack of Platform Accountability
No legislation forcing digital platforms to adopt strict ad transparency.
Inadequate Penalties
No severe legal consequences for hosting fraudulent AI-generated ads.
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.
