Google’s Duplex AI Now Open to Beta Testing

The eerily human-sounding Duplex AI is now in its beta testing phase. The technology sparked a ton of interest when it was introduced at Google’s I/O 2018 conference. On the one hand, this kind of human-sounding virtual assistant was surely the next step for the technology.  

However, concerns were raised over the ethics of the software. It sounds nearly indistinguishable from a human to the untrained ear and was made to make phone calls. Googe says the software will identify itself as non-human when it goes live, but that doesn’t mean there’s no room for misuse. ting phase. The technology sparked a ton of interest when it was introduced at Google’s I/O 2018 conference. 

The Beta Testing 

Google has begun the beta testing phase by having Duplex call businesses that have opted in to the program. Such businesses will receive calls from Duplex to verify things like business hours, and holiday hours. The primary function of the AI, booking appointments, won’t go live until later this summer. For the time being, Google is simply fine-tuning the technology to sound more lifelike and hone its conversational skills.  

A Question of Responsibility 

What kind of responsibility does Google bear in regard to Duplex’s ability to fool people? Technology like this could be used to make a nonliving entity fool humans into thinking it’s alive. Google has attempted to address these fears, stating that the AI will tell whoever it talks to exactly what it is. Something like “Hi, I’m the Google Assistant, calling to make a reservation for a client,” will be stated, perhaps not in those exact words. Even if it’s supposed identify itself as AI, it’s the existence of the technology itself that worries some analysts. 

Google, for its part, has released its manifesto on AI, detailing what they will and won’t develop. It seems they are interested in ensuring the launch of Duplex is ethical. This is good, as it will set the tone for future mass-market AI releases. By setting the right tone out of the gate, Google may be seeking to curb fears that AI will be mishandled. We’ve got our fingers crossed that they know what they’re doing.