Deepfake Elon Musk Promised to Double Crypto, Coinbase Blocked $4.7M
Coinbase says its automated fraud-prevention systems stopped millions of dollars in attempted transfers after scammers used an AI-generated deepfake of Elon Musk to promote a fake Tesla cryptocurrency giveaway.
In a November 17 blog post, published as part of its weekly Consumer Protection Tuesday series, the cryptocurrency exchange described how fraudsters impersonated Musk during a fake livestream of Tesla’s flying-car demonstration.
The broadcast featured a deepfake of Musk promising to double viewers’ crypto deposits. More than 500,000 people watched the livestream, according to Coinbase.
Scammers leveraged a deepfake Elon Musk video during Tesla’s flying‑car livestream to lure users into fraudulent crypto giveaways.
Hundreds of authenticated customers voluntarily transferred funds after being convinced by promises of doubled returns and QR‑code wallet links.
Coinbase’s AI scam‑prevention flagged 91% of attempts by detecting giveaway language, deepfake cues, and new wallet addresses, intervening automatically.
Within approximately two hours, hundreds of Coinbase customers attempted to transfer millions of dollars to the scammers. The exchange said its AI-powered scam-prevention system stopped approximately 91% of the attempted funds, or $4.7 million, from leaving customers’ accounts without human intervention.
However, the figure also highlights the limitations of the intervention. Approximately 9% of the attempted transfers were not intercepted by the system. Coinbase did not disclose the exact amount ultimately received by scammers or independently verified loss figures.
The Scam Used a Fake Tesla Livestream
The fraudsters built the operation around Tesla’s flying-car demonstration, using a YouTube livestream that appeared to resemble the original event.
According to Coinbase, the fake broadcast featured an AI-generated version of Musk, alongside wallet addresses and a QR code. Viewers were promised that any cryptocurrency they sent would be returned at double the original amount.
The exchange described the incident as a social-engineering scam rather than an account compromise. “No Coinbase account was broken into, no password stolen, no credential phished,” the company said.
Instead, hundreds of customers, including longtime platform users, voluntarily initiated transfers after being persuaded by the fake giveaway. “The funds moved because the customer was convinced, not because an attacker forced their way in,” Coinbase explained.
This distinction is important because traditional fraud-detection systems often prioritize unauthorized account access, stolen credentials and suspicious login activity. In this case, customers were authenticated and approved the transfers themselves.
Similar impersonation schemes have previously targeted cryptocurrency users during major technology and space-related events.
The Crypto Times reported in 2023 that scammers used a manipulated Elon Musk video to promote a fake cryptocurrency giveaway, collecting approximately $165,000 from victims who were promised a 200% return.
In another incident in June 2024, scammers used deepfake videos of Musk to promote cryptocurrency giveaways during a SpaceX Starship launch, with fake livestreams appearing across more than 35 YouTube channels.
Note: The Crypto Times is not linking to or embedding the recent Tesla deepfake livestream to avoid directing readers to fraudulent content or inadvertently amplifying the scam.
How Coinbase’s AI Scam-Prevention Agent Worked
Coinbase said the incident exposed a gap in conventional fraud prevention: systems can identify suspicious transaction patterns, but determining whether a customer is being manipulated into authorizing a payment is more difficult.
To address this, the exchange deployed a purpose-built AI scam-prevention agent that combines transaction monitoring with direct customer intervention.
The system has three main components.
- Real-time detection: Machine-learning models assess outgoing transfers and identify risk signals associated with cryptocurrency giveaways, brand impersonation and newly created wallet addresses.
- Customer intervention: When a transfer is flagged, the scam-prevention agent initiates a targeted conversation with the customer. It explains the suspected scam pattern and asks questions intended to help the user recognize inconsistencies in the offer before confirming the transaction.
- Response confidence checks: Coinbase also applies a confidence score to the agent’s responses before displaying them to customers. The company says this is intended to prevent unsupported or unreliable explanations from reaching users.
Coinbase said the agent had not previously encountered this particular scam format, involving a deepfake giveaway linked to a live product demonstration. Nevertheless, it identified several warning signs, including promises to multiply deposited funds, time pressure associated with a live event and the irreversible nature of cryptocurrency transfers.
The exchange said the system operated throughout the incident without human intervention, from the first flagged transfer to the last.
Why AI Deepfakes Complicate Crypto Fraud Prevention
Coinbase argued that generative AI has reduced the resources required to produce convincing impersonation scams.
Previously, creating a convincing fake broadcast could require a production team, recording equipment, and weeks of preparation. AI-generated video and audio can now help scammers reproduce public figures’ appearances and voices with fewer resources. The exchange said this makes fraudulent campaigns easier to launch, modify, and repeat after platforms remove the original content.
Coinbase also outlined additional investments in AI-based risk management, including research-driven detection models, automated agents capable of reviewing different categories of suspicious transfers, and systems designed to classify emerging scam patterns within hours rather than weeks.
However, these capabilities and their reported effectiveness are based on Coinbase’s own account. The exchange did not provide an independent audit of the $4.7 million figure or detailed transaction-level data to establish how much money was ultimately lost.
The episode also illustrates a broader challenge for cryptocurrency platforms: preventing fraud increasingly requires identifying manipulation before a customer authorizes an irreversible transaction, rather than relying exclusively on account-security measures.
What Crypto Users Should Watch For
Coinbase reiterated that legitimate companies and public figures do not ask people to send cryptocurrency first in exchange for receiving a larger amount.
The exchange advised users to avoid sending funds to wallet addresses or QR codes displayed in livestreams, videos, or social media posts. Such addresses can be inserted into content by anyone and do not establish that a promotion is legitimate. It also recommended verifying announcements through official company websites and social media accounts rather than relying on links shared during broadcasts.
Coinbase has published separate guidance on identifying cryptocurrency giveaway scams, warning that fraudsters frequently impersonate companies and executives to make fraudulent promotions appear authentic.
The exchange also advised customers to pause when a transaction triggers a fraud warning and review the reasons before approving it.
Coinbase offered a blunt warning: “We promise: the richest man in the world is not calling you to double your crypto.”
The Bottom Line
The Tesla livestream incident demonstrates how AI-generated impersonation can turn familiar cryptocurrency giveaway scams into more convincing social-engineering campaigns.
Coinbase says its automated defenses prevented $4.7 million in attempted transfers, representing approximately 91% of the funds customers tried to send during the incident. However, the company has not provided independently verified figures for the remaining transfers or confirmed how much ultimately reached the fraudsters.
The case also highlights a shift in cryptocurrency security, with exchanges increasingly using AI to intervene when customers are about to authorize suspicious transactions themselves.
For users, the central warning remains straightforward: a familiar face, a live broadcast or a convincing AI-generated voice does not make a cryptocurrency giveaway legitimate. Any promise to double funds in exchange for an initial crypto transfer should be treated as a scam.
Also Read: CertiK Finds Hidden Hardhat Malware Targeting Crypto Developers’ Keys
Disclaimer: The information researched and reported by The Crypto Times is for informational purposes only and is not a substitute for professional financial advice. Investing in crypto assets involves significant risk due to market volatility. Always Do Your Own Research (DYOR) and consult with a qualified Financial Advisor before making any investment decisions.


