They used synthetic audio in what could be the first major assault carried out by artificial intelligence (AI). They imitated the voice of the executive director of a company and engage in phone conversation with an employee to transfer more than $ 240,000 to a secret account.
The company manager received a night call in which he heard the voice of his superior demanding a transfer to a Hungarian account to save "late payment fines." They sent the financial details by email while talking on the phone. A spokeswoman for Euler Hermes, insurer of the company victim of the theft, said "the software was able to imitate voice, intonation, punctuation and even accent."
The thieves behind the voice called again to demand a second payment, that was when the manager started it to suspect and led him to call his boss. In an email to Euler Hermes, the director said that "a Cyber Johannes demanded to talk to me while I was still on the phone with the real one."
In recent years, hackers and programmers have returned to the so-called deepfakes. Internet platforms cannot detect them and companies struggle on how to handle the consequences.
The constant evolution of deepfakes means that detecting them will never be enough due to the nature of the modern Internet. Last June, Mark Zuckerberg's deepfakes were posted on Instagram shortly after Facebook refused to remove a manipulated video from Nancy Pelosi. There is still no clear consensus on how Facebook should have acted in that situation or in future situations.
This is exaggerated with the data monetization models of companies such as Facebook and Google. Technosociologist Zeynep Tufecki warns that companies like Facebook are confident in creating an "architecture of persuasion" that "makes us more flexible to advertisements, while also organizing our flows of political, personal and social information."
That central dynamic, combined with the constant evolution of deepfake technology, means this problem will worsen on all Internet platforms unless the companies behind them can convince themselves to change their business models.