A fraudster has used an AI-assisted deepfake to scam a company into wiring money to a third-party
The CEO of an energy firm based in the UK sent €220,000 to a third-party at the request of who they believed to be their superior.
Speaking to the Wall Street Journal, Rüdiger Kirsch – a fraud expert at the unknown company’s insurance group – said that the victim had been fooled by the voice, believing it to be that of their boss owing to his German accent and distinctive “melody”.
The caller, who had tried the company three times, instructed the CEO to send the money within the hour to the bank account of their Hungarian supplier. It is believed that the money was then moved to an account in Mexico, then distributed to multiple locations thereafter.
Deepfakes are becoming hard to tell apart
AI-generated deepfakes are becoming harder to differentiate from true speech as the technology becomes more widely available. Last month, Facebook and Twitter were encouraged to combat deepfake videos in the run-up to next year’s US elections.
With a lot of concern over deepfakes becoming the new fake news, some AI companies have displayed their ability to create eerily realistic deepfakes and AI-assisted speech synthesis, not only to entertain but also to warn people of the dangers.
Dessa, an AI company based in Toronto, recently displayed its capabilities by sharing details of their speech synthesis model of Joe Rogan, a popular podcaster and sports commentator.
To mimic Joe Rogan’s voice, Dessa created a text-to-speech deep learning system called RealTalk, which uses text inputs to generate speech indistinguishable from reality.
“Hey Joe Rogan, it’s me, Joe Rogan. Please come save me, man. These artificial intelligence guys have trapped me in a machine,” the AI says in the video, perfectly mimicking Joe Rogan’s distinctive voice and intonation.
In a blogpost, Dessa’s principal machine learning architect, Alex Krizhevksy, warned that “human-like speech synthesis is soon going to be a reality everywhere.”
“The societal implications for technologies like speech synthesis are massive. And the implications will affect everyone. Poor consumers and rich consumers. Enterprises and governments.”