Certified into the record
The exact text we saved
This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.
TRUVACE RECORD VERSION record: TRV-2026-3050 version: 1 kind: certified reason: Certified into the record timestamp: 2026-11-19T12:02:06.261523Z status: published lens: trace sector: crime headline: Coinbase AI Blocked $4.7M In Deepfake Musk Scam dek: Coinbase said its AI scam-prevention agent stopped about $4.7 million, or 91% of attempted sends, from reaching a deepfake Elon Musk crypto giveaway scam. Coinbase (COIN) said its AI scam-prevention agent stopped about $4.7 million in customer funds from reaching scammers who used a deepfake of Elon Musk to pitch a fake Tesla crypto giveaway, blocking roughly 91% of the money customers tried to send. T gain_title: Coinbase said its AI scam-prevention agent blocked about $4.7 million, roughly 91% of the money customers tried to send, to a deepfake Elon Musk crypto giveaway scam, with no humans in the loop. problem_title: Coinbase's figures imply about $460,000 still left customers' accounts during the deepfake Musk giveaway scam, even with its AI agent running, and the company did not say how much reached the scammers. trace_subject: Coinbase AI agent against the deepfake Musk giveaway scam gain_reading: Coinbase said its AI scam-prevention agent blocked about $4.7 million, roughly 91% of the money customers tried to send, to a deepfake Elon Musk crypto giveaway scam, with no humans in the loop. gain_evidence: stopped about $4.7 million in customer funds from reaching scammers | blocking roughly 91% of the money customers tried to send | with no humans in the loop problem_reading: Coinbase's figures imply about $460,000 still left customers' accounts during the deepfake Musk giveaway scam, even with its AI agent running, and the company did not say how much reached the scammers. problem_evidence: Those two figures imply customers tried to send about $5.2 million, and that roughly $460,000 still left their accounts | Coinbase did not say how much reached the scammers quick_read: Coinbase said its AI scam-prevention agent stopped about $4.7 million in customer funds from reaching scammers who used a deepfake of Elon Musk to promote a fake Tesla crypto giveaway. That was roughly 91% of the money customers tried to send. The company described the incident in a Tuesday blog post and said the system worked with no humans in the loop. The scam was a YouTube livestream timed to Tesla's flying car demo, with a deepfake Musk, on-screen wallet addresses and a QR code promising to double any crypto sent. More than half a million people watched, Coinbase said, and hundreds of its customers tried to send millions of dollars over about two hours. The two figures imply customers tried to send about $5.2 million and that roughly $460,000 still left their accounts. Coinbase did not say how much reached the scammers, when the stream ran or which coins customers tried to send. The numbers are the company's own account of its own system. limitation: The $4.7 million and 91% figures come from Coinbase's own blog post, and the company did not say how much reached the scammers, when the fake stream ran or which coins customers tried to send. tag: Dual reading key_points: Coinbase said the agent stopped about $4.7 million, roughly 91% of attempted sends, so about $5.2 million was implied to have been attempted. | Roughly $460,000 still left customer accounts, by the article's arithmetic from Coinbase's figures. | The fake livestream ran during Tesla's flying car demo and drew more than half a million viewers, Coinbase said. | Over about two hours, hundreds of Coinbase customers, many of them long-time users, tried to send millions of dollars to the on-screen addresses. | Coinbase said no account was hacked; customers logged in and confirmed the sends themselves, which makes giveaway fraud hard to stop. | The agent had never been trained on deepfake giveaways tied to a live launch, but recognised the promise to multiply funds, the live-event deadline and an irreversible transfer. | Coinbase is expanding AI review to every type of scam and using AI to label new cases so models learn from attacks within hours instead of weeks. rundown: Coinbase says its defense has three parts. Machine learning models score outgoing sends in real time, looking for giveaway patterns, brand impersonation and newly created wallet addresses. When a send is flagged, a scam-specific agent looks up context, names the scam pattern, explains how giveaway schemes work and asks questions meant to make the story fall apart. Each reply is scored for confidence before the customer sees it, so the system knows when more friction is needed. The company stressed that its platform was not breached: no account was broken into, no password stolen, no credential phished. Because customers authorised the transfers themselves, defenses that look for unauthorised activity do not catch this fraud. Coinbase wrote that its systems can see risk signals in real time but cannot see intent. Coinbase said deepfakes have changed the economics of such scams: staging one once took a crew, a studio and weeks, and now takes a livestream slot and a model. The post was part of a weekly consumer-protection series. Coinbase closed it with a list of things it will never do, including promising to double, triple or multiply crypto that customers send. sources: - journalism | Bitcoin Magazine | https://bitcoinmagazine.com/news/coinbase-ai-blocks-deepfake-musk-scam | 2026-11-18 prev: 0000000000000000000000000000000000000000000000000000000000000000
- SHA-256 hash
- 14799cd5652d28f1fa3c14251719a72dbe53b8a6f75994c46ed97dae7684d6d2
- Previous version’s hash
- 0000000000000000000000000000000000000000000000000000000000000000
How to check it yourself, without trusting this page
Download any version from /api/record/TRV-2026-3050 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.
ace