We can’t stop progress… and the bad people understand that. If theartificial intelligence opens the way to great advances, it also provides new tools to perfect scams.
Sam Mitrovic, a cybersecurity expert, paid the price this summer and recounted it on his blog, spotted this Sunday, October 13 by French media.
Email then phone call
He first receives an email attempting to recovering your Gmail accountthen about 45 minutes later a phone call appearing to be from Google, which he ignores.
A week later, do it again. This time, Sam Mitrovic wins. At the end of the line, “an American voice very polite and professional. The number is Australian”says the expert.
His interlocutor alerts him to suspicious activity, sends him a summary email at his request. Everything seems true. Except nothing is.
“Super realistic” voice
The voice on the phone, although “super realistic”was in fact an artificial intelligence. Sam Mitrovic realized this thanks to pronunciation and cadence “too perfect”.
“If I had stayed on the phone long enough, the next step would have been to approve the recovery of my account, which would have given them access.”
The sender of the email seemed to come from the google.com domain but it was a manipulation: among the recipients, there was the real address which was googlemail[@]internalcasetracking.com.
Among the other flaws in this attempted cyberattack, the expert noted that he had not initiated a request for account recovery, that Google did not call personal accounts, that there was no suspicious connection to his account and that the phone number was notified in several forums as being a scam.
“THE scams are becoming more and more sophisticatedconvincing and are deployed on an ever-larger scale”alert Sam Mitrovic. “That one seemed totally legit, I give it an A for effort. A lot of people might fall for this trick.”
