Scammers use AI-generated avatars to impersonate friends on Facebook and solicit money
Scammers in Houston, Texas reportedly used AI-generated deepfake videos to impersonate Facebook users and their friends, extracting access codes and soliciting money for fake items for sale. Multiple victims reported losing money after being deceived by deepfake videos that mimicked voices, faces, and mannerisms of people they knew.
What happened
In Houston, Texas, scammers used AI-generated avatars to impersonate friends and solicit money from victims. Two Houston-area women, Stacey Svegliato and Sara Sandlin, reported being victimized after receiving video calls and messages that appeared to be from acquaintances on Facebook Messenger.
According to the victims, scammers extracted access codes by impersonating hacked friends, then posted items for sale on the victims' Facebook pages using AI-generated videos. The deepfake videos were sophisticated enough to mimic facial expressions, voice, and mannerisms, leading some of the victims' friends to send hundreds of dollars for non-existent items.
Reported impact
According to the victims, some of their friends sent hundreds of dollars to scammers posing as the AI-generated avatars, believing they were communicating with the impersonated friends. Sara Sandlin stated that 'some people fell for it, some people didn't,' indicating that the deepfake videos successfully deceived multiple individuals.
Index analysis
Cybersecurity expert Dominic Chorafakis noted that scammers are using videos posted online to create AI-generated avatars that can fool people who know the impersonated individuals well. According to Chorafakis, the technology only requires a still image or short video segment, with lip-syncing handling the rest of the video generation.
Governance controls
Chorafakis recommended limiting what personal content is shared publicly and using platform security features to restrict access to videos and images. The victims expressed concern that scammers could continue using their avatars for future scams.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.