Scammers attempted to defraud a Veriff employee by using a real-time deepfake video call to impersonate the company's CEO, Kaarel Kotkas. The employee identified the fraud due to inconsistencies in the CEO's accent and verified the attempt via an internal messaging platform. No financial loss occurred, but the incident highlights the increasing sophistication of real-time synthetic media in social engineering attacks.
Scammers reportedly targeted Veriff colleague Andrea Rozenberg through WhatsApp with an urgent request and a video call that appeared to show Veriff CEO Kaarel Kotkas. Kotkas and Veriff later described the call as a real-time deepfake or live video clone. Rozenberg reportedly noticed the missing Estonian accent and verified through Slack that the request was fraudulent. No financial loss was reported.
Risk classification
- Primary risk domain: 4 Malicious actors
- Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation
The incident involved scammers using real-time deepfake video and voice cloning technology to impersonate a CEO in a targeted attempt to defraud an employee.
Causal factors
- Entity: Human
- Intent: Intentional
- Timing: Post-deployment
The incident was caused by human scammers intentionally using deployed AI deepfake tools to execute a social engineering attack.
EU AI Act risk tier
- Risk tier: 3 Limited Risk
Risk Level 3: Limited Risk. The report describes the use of 'AI-generated content such as deepfakes' to impersonate the CEO.
AI system and alleged parties
- AI system: unspecified
- AI purpose: Deepfake Video Generation; Voice Generation
- Behaviour type: Tool
- Alleged developer: Deepfake technology developers
- Alleged deployer: Scammers
- Alleged harmed parties: Victims of deepfake abuse, Veriff, Kaarel Kotkas, Epistemic integrity, Andrea Rozenberg
Harm severity
Severity ratings are based on this scale.
Highest direct severity in any category: Negligible. 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 Negligible, indirect Negligible
- Environmental: direct Negligible, indirect Negligible
- Malicious content: direct Negligible, indirect Negligible
- Differential treatment: direct Negligible, indirect Negligible
- Civil rights: direct Negligible, indirect Negligible
- Democracy: direct Negligible, indirect Negligible
- Privacy: direct Negligible, indirect Negligible
- Psychological: direct Negligible, indirect Negligible
- Epistemic: direct Negligible, indirect Negligible
- Child sexual exploitation and abuse: direct Negligible, indirect Negligible
People affected
- Occurrences reported: 1
- People reportedly exposed: 2
Potential causes
Management
- Outdated Security Frameworks: Relying on checkbox compliance instead of proving true human presence.
Technology
- Real-time Deepfake Technology: Free tools enable scammers to fake video streams in real time.
- Imperfect Acoustic Cloning: The AI tool failed to replicate the target's Estonian accent.
Data Inputs
- Inadequate Accent Training Data: The voice clone lacked specific training data on Estonian accents.
Human Factors
- Targeted Social Engineering: Attackers targeted a specific executive on a highly relevant topic.
- Reliance on Human Detection: Security relied on individuals detecting subtle anomalies in video calls.
Process and Methods
- Out-of-band Verification Gap: No automated system to verify identity before accepting executive calls.
Regulatory Environment
- Unregulated AI Cloning Tools: Easy access to free voice and video cloning tools online.
Information quality
- Classification confidence: High
- Reason for confidence: The report is a first-hand account from the impersonated CEO, clearly describing the event, how it was detected, and the outcome. There is no conflicting information.
An attempted corporate fraud utilizing a real-time deepfake video and voice clone of a CEO was successfully thwarted by a vigilant employee. The incident demonstrates the growing sophistication of consumer-accessible AI tools for social engineering, but poses negligible direct threat to national security.
- Overall national security impact: Minor
- Scope: Unknown
- Primary target: No clear primary
- Alleged perpetrator: Unknown
Threat characteristics
- Imminence: Long-term. Represents an ongoing strategic cyber-hygiene and corporate security concern rather than an active national security crisis.
- Autonomy: Human-controlled. The AI tools were directly operated and controlled by human scammers to execute a social engineering attack.
- Novelty: Evolved capability. Represents an evolution of existing social engineering and deepfake capabilities, applying real-time video generation in a live call.
Impact by dimension
- Physical security: Negligible. No physical systems, infrastructure, or human safety was threatened or impacted by this commercial fraud attempt.
- Information security: Negligible. No state-sponsored disinformation, intelligence compromise, or mass surveillance capabilities were involved in this localized commercial fraud attempt.
- Sovereignty: Negligible. No government processes, elections, or sovereign state decisions were targeted or compromised in this incident.
- Economic security: Minor. The incident represents a minor corporate security threat using voice and video cloning, but resulted in no financial loss or compromise of strategic technology.
- Societal stability: Negligible. The incident was a targeted corporate scam and did not affect societal stability, civil liberties, or the general public.
.avif){kind=small}