This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Scammers are using highly personalized, AI-generated invitation emails to impersonate famous podcasters and defraud entertainment industry figures of money and social media account credentials.
Broadway performer and SiriusXM host Seth Rudetsky reportedly received a personalized invitation purportedly from comedian Nikki Glaser's podcast. After scammers guided him through Facebook permissions during a supposed technical setup, he lost access to his page; he said a former marketing company was also compromised. Reporting described the invitation as most likely AI-composed.
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 involves malicious actors using AI-generated text to impersonate famous podcasters in highly targeted phishing scams to defraud victims of money and hijack social media accounts.
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 caused by human malicious actors intentionally using post-deployment AI text generation tools to draft highly persuasive phishing emails.
A localized phishing campaign utilized AI-generated text to impersonate famous podcasters and defraud entertainment industry figures in the United States. While demonstrating the evolving capability of AI-assisted social engineering, the incident remains a standard cybercrime matter with negligible implications for national security.
The report describes scammers using AI as a writing assistant to compose 'flattering, highly personalized invitations' and 'imitate Terry Gross' voice itself'. The AI system does not act autonomously to execute the scam; it merely assists the human scammers by generating highly persuasive text.
Risk Level 3: Limited Risk. The system is described as generating AI-generated text content that impersonates real individuals ('imitate Terry Gross' voice itself', 'AI themselves into all of the work'), which falls under the category of AI-generated content requiring transparency.
The AI was used to draft highly personalized, articulate, and persuasive invitation emails that replicated the writing style and voice of specific famous podcasters.
The report provides detailed first-hand accounts from two prominent victims of the scam, outlining the exact methods used by the fraudsters. While the use of AI is described as 'most likely' rather than definitively proven by technical forensics, the highly personalized and articulate nature of the text strongly aligns with modern LLM capabilities, and the victims themselves attribute the feasibility of the scam to AI.
The emails could have been written entirely by highly dedicated human scammers without AI assistance, though the scale and specificity make AI highly probable.
Technical confirmation of the specific AI tool used and forensic headers of the emails.
Technology
AI-Generated Persuasive Text
AI generated highly articulate emails mimicking 'Terry Gross' voice.
Lack of Email Authentication
No robust email verification allowed scammers to spoof booking requests.
Data Inputs
Harvested Career Details
AI used specific public data like Broadway cruises to build false trust.
Human Factors
Cognitive Bias of Flattery
Victims believed the scam due to highly personalized and flattering invites.
Unfamiliarity with AI Risks
Victims did not consider AI capabilities to automate deep research.
Benefit of the Doubt
Victims naturally trusted the scammers, overlooking initial red flags.
Process and Methods
Flawed Booking Verification
Lack of standard protocol to verify podcast invitations independently.
Insecure Permission Delegation
Sharing Facebook admin controls during a Zoom call without verification.
Management
Weak Credential Governance
Marketing firm retained legacy Facebook admin rights, exposing the page.
Regulatory Environment
Lack of Deepfake Regulations
Absence of strict laws governing AI-powered text and voice impersonation.
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.
