Coinbase said its AI-powered scam detection system stopped $4.7 million, which is around 91% of the funds targeted by scammers, from reaching their wallets during a fake Elon Musk crypto giveaway.
What makes the incident stand out is that the entire operation was handled without human intervention, as the company’s AI agent identified the scam and stepped in as customers attempted to send their money.
The scheme revolved around a fraudulent YouTube livestream that appeared to show Tesla’s flying car demonstration.
Deepfake Crypto Scam
The fake livestream used an AI-generated version of Elon Musk, displayed crypto wallet addresses and a QR code, and promised viewers double their money if they sent cryptocurrency. More than 500,000 people watched the stream, giving the fake event plenty of exposure. According to Coinbase, the difficult part was that the scammers did not need to break into anyone’s account or steal login details.
Customers were logged in, had access to their funds, and willingly approved the transfers because they believed the giveaway was legitimate.
That makes this type of fraud harder to catch using traditional security measures, which generally focus on unauthorized account access or suspicious activity.
To address the problem, Coinbase deployed an AI scam-prevention agent designed to flag risky transfers and challenge customers before their money leaves the platform. The system uses machine learning models to assess transactions in real time. It looks for warning signs such as giveaway-related activity, brand impersonation, and transfers to newly created wallet addresses.
When a payment is flagged, the AI agent steps in with a targeted conversation that explains the specific scam pattern and why the offer should raise concerns, rather than relying on a generic security warning. Coinbase also said that every response generated by the agent receives a confidence score before being shown to the customer, which helps the system determine when it has enough information to intervene confidently.
Interestingly, the said crypto exchange agent had never encountered this particular scam before. Deepfake giveaways connected to live product launches were not part of its training, but it still recognized familiar warning signs, including promises to multiply crypto holdings, time pressure, and irreversible transfers. The system reportedly handled the flagged attempts from the first transaction to the last without requiring a human to step in.
“Before, staging a scam like this took a crew, a studio, and weeks. Now it takes a livestream slot and a model. Deepfakes make convincing scams cheap, fast, and easy to relaunch under a new name the moment one is shut down. That means our defenses have to improve at the same rate the attacks do. We’re investing heavily in AI across Risk to make that happen.”
Major Attack Vector
A Gartner survey of 297 senior cybersecurity leaders, conducted between March and May 2026 and published in September, found that 41% of CISOs had dealt with at least one social engineering incident involving an employee audio deepfake over the previous 12 months. Video deepfakes weren’t far behind. More than one in three CISOs, or 36%, said their organization had experienced this type of incident.
In June, Japanese crypto advocate Mai Fujimoto was tricked by a deepfake impersonating an acquaintance during a Zoom call, which eventually led to malware stealing access to her accounts. Binance founder CZ later warned that even video calls can no longer guarantee someone’s identity.
Le poste Coinbase AI Thwarts Stops Elon Musk Crypto Scam, Blocking Millions of Stolen Funds est apparu en premier sur CryptoPatate.














Laisser un commentaire