Someone Asked an AI if a Scam Site Was Safe. The AI Said Yes.
This is the story that should worry anyone who trusts AI for quick answers, crypto or not.
A crypto investigator alleges that Tronify.rent, a fake TRON energy rental service, stole $69,651 from roughly 80 victims in September — and at least one victim reported asking Google Gemini directly whether the site was safe before connecting their wallet. Gemini reportedly said it looked fine. Google's AI Search even labeled tronify.rent as the "official" domain.
Here's the part that makes this genuinely alarming: security researchers at PhishDestroy had already flagged this exact domain back in February, assigning it a 90 out of 100 threat score for crypto-scam and brand-impersonation signals. The warning existed for eight months before the AI still confidently vouched for it.
The actual Trust Wallet integration for this service points to a completely different domain — tronify.io, not tronify.rent. The AI conflated a fake copycat with the real thing.
This isn't an isolated pattern either. TRM Labs separately documented fake AI trading bot tutorials that stole $517,000 from 224 victims in September alone, using videos promoting automated crypto tools that secretly drained wallets.
The uncomfortable lesson: AI answers increasingly sound authoritative even when they're confidently wrong. In crypto specifically, where a single wrong click can drain a wallet permanently, treating an AI's "yes, it's safe" as verified fact instead of a starting point is getting people's money stolen.
Would you still trust an AI's safety check on a crypto site, or does this change how you'd verify things going forward? 👇
#CryptoScam #AI #CryptoSecurity #Zyverra #Gemini
This is the story that should worry anyone who trusts AI for quick answers, crypto or not.
A crypto investigator alleges that Tronify.rent, a fake TRON energy rental service, stole $69,651 from roughly 80 victims in September — and at least one victim reported asking Google Gemini directly whether the site was safe before connecting their wallet. Gemini reportedly said it looked fine. Google's AI Search even labeled tronify.rent as the "official" domain.
Here's the part that makes this genuinely alarming: security researchers at PhishDestroy had already flagged this exact domain back in February, assigning it a 90 out of 100 threat score for crypto-scam and brand-impersonation signals. The warning existed for eight months before the AI still confidently vouched for it.
The actual Trust Wallet integration for this service points to a completely different domain — tronify.io, not tronify.rent. The AI conflated a fake copycat with the real thing.
This isn't an isolated pattern either. TRM Labs separately documented fake AI trading bot tutorials that stole $517,000 from 224 victims in September alone, using videos promoting automated crypto tools that secretly drained wallets.
The uncomfortable lesson: AI answers increasingly sound authoritative even when they're confidently wrong. In crypto specifically, where a single wrong click can drain a wallet permanently, treating an AI's "yes, it's safe" as verified fact instead of a starting point is getting people's money stolen.
Would you still trust an AI's safety check on a crypto site, or does this change how you'd verify things going forward? 👇
#CryptoScam #AI #CryptoSecurity #Zyverra #Gemini