AI stops Tesla scam before victims lose millions
Fighting fire with fire.

The Tesla brand logo in Parsdorf near Munich, Germany. Matthias Balk/picture alliance via Getty.
- Coinbase says AI defenses blocked about 91% of targeted funds in a Tesla-themed scam.
- Scammers impersonated Tesla and Elon Musk during a fake YouTube livestream promoting a crypto giveaway.
- More than 500,000 viewers watched the stream, and Coinbase users tried to send millions within two hours.
- Coinbase says its AI agents warned users in real time without employee intervention.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
In an intensifying tug-of-war between AI-assisted criminals and AI-powered defenders, cybersecurity professionals used AI to thwart a popular Tesla scam.
Major crypto exchange Coinbase said criminals attempted to impersonate Tesla and its boss, Elon Musk, to steal funds from the exchange's users.
On YouTube, scammers promoted a fake crypto asset giveaway during Tesla's "flying car" demonstration in October. According to Coinbase, the "livestream" managed to attract more than 500,000 viewers, while hundreds of the exchange's customers attempted to send millions of dollars to the scammers in around two hours.
These potential victims were tricked by promises to double their deposits if they sent crypto assets to a provided address, which viewers didn't know was controlled by scammers.
However, Coinbase says its AI-powered defenses helped prevent around 91% of the targeted money, or $4.7 million, from leaving customers' accounts, even though these customers authorized the transactions themselves.
The exchange detailed that its AI agents evaluated transactions in real time, looking for suspicious patterns associated with crypto giveaways, brand impersonation and newly created wallet addresses.
Next, when a potentially fraudulent transaction was flagged, an AI agent contacted the affected customer, explaining the specific scam pattern and why the transaction was dangerous.
According to the company, the entire process was handled without employee intervention.
Coinbase emphasized that its defenses should improve at the same rate as attacks, meaning AI can also help identify new scam formats.
Stay updated with our latest stories and follow us on social media
Be the first to discover new stories, ideas, and updates from our team.
"Models get better through AI-driven research, so detection keeps pace with new scam formats instead of waiting for the next wave," the company said, adding that its models "learn from each attack sooner – often within hours instead of weeks."
This is not the first time crypto scammers have tried to hijack popular livestreams. As reported by Cybernews, in October 2025, criminals targeted Jensen Huang, then-CEO of Nvidia, the world's most valuable company. They hijacked a livestream and promoted a crypto scam, managing to attract more than 100,000 viewers.