Coinbase blocks Musk deepfake scam transfers with AI defenses

Coinbase blocks Musk deepfake scam transfers with AI defenses
Coinbase halts deepfake scam

Crypto scams using celebrity impersonation continue to test platforms' ability to stop user-authorized transfers before funds leave customer accounts. Coinbase says its AI-based defenses halted $4.7 million in attempted payments tied to a fake Tesla livestream that used a deepfake of Elon Musk to promote a bogus crypto giveaway.

Highlights

  • During Tesla's 'flying car' livestream, scammers' deepfake Elon Musk giveaway drew more than 500,000 viewers and prompted millions in attempted crypto transfers.
  • Coinbase's AI defense system automatically blocked about 91% of fraudulent customer-initiated transactions to scam wallet addresses within two hours, despite users themselves authorizing the transfers.
  • Coinbase plans to increase investment in AI fraud prevention, adapting machine-learning models to detect new scam formats like deepfake live event impersonations and evolving giveaway tactics.

Real-time scam detection during Tesla livestream

As reported by The Block, citing Coinbase in a blog post published Tuesday, the attempted fraud took place during Tesla's "flying car" demonstration last month, when scammers launched a YouTube stream designed to look like the automaker's official broadcast.

The fake stream featured a deepfake of Elon Musk, along with wallet addresses and a QR code promising to double any cryptocurrency sent. Coinbase says more than 500,000 people watched the livestream, and within about two hours, hundreds of its customers tried to send millions of dollars to the displayed addresses.

The company says its system prevented about 91% of the attempted transfers without human intervention. Coinbase adds that the case was especially difficult because customer accounts were not hacked and users voluntarily authorized the transfers after being misled by the scam.

AI defense investment targets new scam patterns

Coinbase says conventional real-time fraud tools are less effective when users themselves approve a transaction, even if they have been deceived. The company says its AI defense system addresses that gap by using machine-learning models to score risky transfers, including activity matching giveaway scams, brand impersonation and newly created wallet addresses.

After flagging the activity, Coinbase says a dedicated scam-prevention agent assessed the transaction context, explained how such scams typically work and asked questions intended to break down the fraud narrative. The company says the agent had not previously encountered this specific format, a deepfake giveaway tied to a live product event, but still identified the core warning signs, including promises to multiply funds, time pressure and irreversible transfers.

Coinbase says it will keep investing heavily in AI defenses as it seeks to adapt its models to emerging scam techniques through AI-driven research.

The latest update follows our earlier article on crypto security gaps as AI tools heighten risks for developers.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.