This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Fraudsters exploited the timing of Apple's iPhone 14 announcement to host fake YouTube live streams. These streams used repurposed, edited footage of tech CEOs like Tim Cook and Elon Musk to deceive viewers into visiting fraudulent cryptocurrency websites. The streams reached tens of thousands of viewers before being removed by YouTube for violating terms of service.
Fraudsters repurposed an old interview with Apple CEO Tim Cook using AI or video editing to promote a fake crypto event on YouTube. The altered video was designed to mislead viewers into believing Tim Cook endorsed a new cryptocurrency scheme. The stream attracted tens of thousands of viewers, potentially exposing them to the scam before it was removed for violating YouTube’s terms of service.
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 fake streams spread misleading information about Apple launching a cryptocurrency venture, which was amplified by YouTube's recommendation algorithm.
The incident involves fraudsters using the public likenesses of tech CEOs in fake live streams to run a cryptocurrency scam, which falls under fraud, scams, and targeted manipulation.
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 fraudsters who intentionally hijacked YouTube channels and edited videos of tech executives to run a cryptocurrency scam post-deployment.
Added categories peak at 3: Substantial (Epistemic Harm, Inferred)
Fraudsters hijacked YouTube channels to broadcast fake live streams promoting a cryptocurrency scam, using repurposed footage of prominent tech CEOs. While the streams reached over 80,000 viewers before being removed, the incident represents a low-level financial scam with negligible direct national security implications.
The reports do not identify or characterize any specific AI system's mode of operation, focusing instead on the human fraudsters' actions and the general platform recommendation mechanism.
Risk Level 3. Limited Risk: The incident involves the use of manipulated media and deepfakes/repurposed video to deceive users, which falls under transparency obligations for AI-generated content. The report notes that 'Fraudsters take an image of a famous person... and package videos... to make it appear like they're talking about some new crypto venture.'
The primary AI system mentioned is YouTube's recommendation algorithm, which promoted the scam stream to users. The secondary goal relates to the potential use of video manipulation tools to package the fake streams.
The reports clearly describe the mechanics of the scam, the channels hijacked, the specific videos repurposed, and the viewer counts. There is little ambiguity about the nature of the event, although the specific AI tools used to edit the videos (if any) are not detailed.
None. The event is clearly a coordinated cryptocurrency scam exploiting a major tech announcement.
While details on the specific video editing tools or the exact financial losses of victims are absent, they are not strictly necessary to classify the incident as a malicious scam.
Technology
Recommendation Algorithm Bias
Algorithm recommended scam streams based on user's recent watch history.
Inadequate Automated Moderation
Automated systems failed to detect and block the fraudulent live stream.
Data Inputs
Manipulated Stream Metadata
Scammers filled titles and descriptions with trending Apple keywords.
Hijacked Channel Data
Algorithm used historical data from hijacked accounts to trust the stream.
Human Factors
User Trust in Recommendations
Viewers assumed recommended homepage streams were safe and legitimate.
Exploitation of Event Buzz
Scammers targeted users looking for Apple's Far Out event announcements.
Process and Methods
Inadequate Stream Verification
No proactive verification of high-traffic live streams before promotion.
Management
Engagement-Driven Optimization
Platform prioritized viewer engagement metrics over content safety checks.
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.
