This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A deepfake video and audio of former FTX CEO Sam Bankman-Fried was used in a scam on Twitter targeting victims of the FTX exchange bankruptcy. The content falsely promised to double cryptocurrency holdings as compensation for losses, directing users to a fraudulent website. The scam resulted in a small amount of financial loss before the account was suspended.
A visual and audio deepfake of former FTX CEO Sam Bankman-Fried was posted on Twitter to scam victims of the exchange's collapse by urging people to transfer funds into an anonymous cryptocurrency wallet.
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 generated entirely fabricated statements claiming that FTX was hosting a cryptocurrency giveaway to compensate users.
The incident involves the use of a deepfake to impersonate a public figure for the purpose of defrauding victims of the FTX collapse.
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 a human actor who intentionally created and deployed a deepfake of Sam Bankman-Fried to execute a financial scam.
Added categories peak at 2: Minor (Epistemic Harm, Inferred)
A deepfake video of former FTX CEO Sam Bankman-Fried was used in a minor cryptocurrency scam on Twitter. The incident resulted in negligible financial loss and poses no threat to national security, representing an established cybercrime method rather than a state-level threat.
The AI system was used as a tool to generate a deepfake video and audio of Sam Bankman-Fried. It performed a specific, well-defined function of synthesizing media based on direct inputs without autonomous planning or decision-making.
Limited Risk: The report describes the creation of a deepfake, which falls under Risk Level 3 as it requires transparency obligations to inform users they are interacting with AI-generated content.
The AI was used to generate a convincing synthetic video and audio (deepfake) of Sam Bankman-Fried to execute a scam.
The report clearly details the nature of the scam, the specific deepfake used, the platform (Twitter), and the financial impact (0.89 ETH). The role of the AI as a generative tool is clear, and the boundary rules easily resolve the classification.
None
Technology
Convincing Deepfake Generation
ML and AI techniques generated realistic visual and audio content of SBF.
Malicious Scam Website
A live phishing website was used to collect cryptocurrency from victims.
Data Inputs
SBF Audio and Video Data
Abundant public media of SBF allowed training of realistic deepfake models.
Human Factors
Victim Desperation for Recovery
FTX collapse victims were desperate to recover their lost assets.
Trust in Twitter Verification
Users trusted the blue checkmark on the malicious account.
Process and Methods
Flawed Twitter Verification Policy
Twitter's paid verification allowed scammers to obtain verified status easily.
Delayed Account Suspension
The scam Twitter account remained active long enough to post the deepfake.
Management
Weak Platform Content Moderation
Twitter management failed to proactively block deepfake scam campaigns.
Inadequate Risk Assessment
Failure to assess risks of paid verification policy exploitation.
Regulatory Environment
Lack of Deepfake Regulation
Absence of strict legal frameworks governing the creation of deepfakes.
Unregulated Crypto Platforms
Lack of oversight on crypto transactions enabled rapid transfer of funds.
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.
