Amazon unveiled its Alexa voice-mimicking technology at this year’s MARS Conference, and this new technology is already bringing up controversy among tech experts. Nima Schei, CEO of computer vision startup Hummingbirds AI was interviewed by NTD Business News to get his expert insights about Alexa’s voice-mimicking technology opening a door to cybercrime and authentication fraud.
Every edition of the MARS Conference, brings together innovative minds in Machine learning, Automation, Robotics, and Space with company product showcasing. In this edition, the new voice-mimicking technology in Alexa was introduced, causing opposing reactions among experts in the field of machine learning but also in cybersecurity. The reason? Hummingbirds AI´s CEO and computer vision expert Nima Schei, explains:
“The risk is so big we don’t even want to think about it. In the near future we will have more and more attacks from people trying to impersonate you through your voice or later through your visuals.” As the CEO and creator of privacy-first vision AI solutions, Nima makes sure all the solutions out of Hummingbirds AI have privacy in their core model, like GuacamoleID, the continuous re-authentication tool for enterprises to maximize productivity and improve their device security by blocking unauthorized access through vision AI and facial biometrics, Nima Schei understands perfectly the implicit risk in this new technology, which is the opening doors to more fraud and impersonations, now through sound and voice.
Amazon voice-mimicking technology hasn’t been released yet but according to Nima Schei, the negative sides are clear: “it could be used by fraudsters or malicious people in order to abuse this kind of technology. There have to be protocols from Amazon and similar companies to limit this type of abuse.” One of the most notorious crimes involving deep fake technology was a 2020 incident in the United Arabs Emirates where fraudsters tricked a bank manager into transferring 35 million dollars.