This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A LastPass employee was targeted by a social engineering attack involving an audio deepfake impersonating the company's CEO via WhatsApp. The employee recognized the suspicious nature of the communication and reported it to the internal security team, resulting in no security breach. LastPass publicized the incident to raise awareness about the increasing use of deepfake technology in corporate fraud campaigns.
A LastPass employee was reportedly targeted by an audio deepfake impersonating CEO Karim Toubba via WhatsApp. The message used voice-cloning AI in a social engineering attempt to create urgency and bypass security protocols. The employee recognized red flags, reported the incident, and no breach occurred. LastPass shared the incident publicly to raise awareness about deepfake-driven fraud campaigns targeting executives and companies.
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.
The incident involved a targeted social engineering scam where fraudsters used an AI-generated voice clone of the LastPass CEO to attempt to deceive an employee for potential financial or operational gain.
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 initiated by a human threat actor who intentionally used generative AI voice-cloning technology post-deployment to execute a social engineering attack.
An attempted social engineering attack against a LastPass employee using an audio deepfake of the CEO was successfully thwarted. The incident resulted in no security compromise, financial loss, or national security impact, representing a localized corporate fraud attempt rather than a threat to national security.
The AI system (voice cloning/generative audio) was used as a tool by a human threat actor to generate a synthetic voice clone. The reports describe the technology as being programmed by the creator to fabricate recordings, operating on direct inputs without autonomous goal-directed planning or interactive assistant capabilities.
Limited Risk: The report describes the use of 'AI-generated content such as deepfakes', which falls under Risk Level 3 due to transparency and awareness obligations regarding synthetic media.
The AI system was used to clone and generate the voice of the LastPass CEO to impersonate him in phone calls and voicemails.
The reports from SC Media, Hackread, and LastPass's official blog provide consistent, clear, and detailed accounts of the specific attempted attack on the LastPass employee. There is no ambiguity about the outcome (thwarted, no breach, no financial loss) or the technology used (audio deepfake/voice cloning via WhatsApp).
None
Technology
Advanced Voice Cloning Tools
Enables scammers to replicate executive voices with high accuracy.
Camera Injection Vulnerabilities
Allows feeding synthetic media directly into video streams.
Accessible Deepfake Platforms
Lowers the skill barrier to create high-quality synthetic media.
Data Inputs
Publicly Available Media Samples
Provides source audio and video of executives to train AI models.
Human Factors
Vulnerability to Social Engineering
Urgency and trust cues trick employees into bypassing protocols.
Remote Work Communication Gaps
Prevents quick face-to-face verification of suspicious requests.
Process and Methods
Use of Unofficial Channels
Conducting critical business communications over apps like WhatsApp.
Weak Financial Approval Workflows
Allows single-point authorization of high-value transactions.
Absence of Liveness Checks
Fails to verify physical presence during video conferences.
Management
Inadequate Employee Training
Fails to prepare staff to identify sophisticated deepfake tactics.
Delayed Policy Updates
Internal protocols fail to keep pace with rapidly evolving AI threats.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
