Google’s virtual assistant can now make phone calls on your behalf to schedule appointments, make reservations in restaurants and get holiday hours.
The robotic assistant uses a very natural speech pattern that includes hesitations and affirmations such as “er” and “mmm-hmm” so that it is extremely difficult to distinguish from an actual human phone call.
The unsettling feature, which will be available to the public later this year, is enabled by a technology called Google Duplex, which can carry out “real world” tasks on the phone, without the other person realising they are talking to a machine. The assistant refers to the person’s calendar to find a suitable time slot and then notifies the user when an appointment is scheduled.
Google’s CEO, Sundar Pichai, demonstrated the capability on stage at the Shoreline Amphitheater during the company’s annual developer conference, I/O. He played a recording of Google Assistant calling and interacting with someone at a hair salon to make an appointment.
“What time are you looking for?” the salon employee asks. “At 12pm,” replies the robot. The salon doesn’t have an opening then, so the robot suggests a window between 10am and 12pm, before confirming the booking and notifying its human master.
From the onstage demonstrations, it seemed like a significant upgrade from the automated phone systems most people have interacted with. The natural interaction is enabled by advancements in automatic speech recognition, text-to-speech synthesis and an understanding of how humans pace their conversations.
Pichai said the tool was useful for the 60% of small businesses in the US that don’t already have an online booking system. “We think AI can help,” he said.
The Duplex system can also call a company to ask about hours of operation during a holiday, and then make that information available online with Google, reducing the volume of similar calls a business might receive.
“Businesses can operate as they always have. There’s no learning curve or changes to make to benefit from this technology,” the principal engineer, Yaniv Leviathan, and vice-president, engineering, Yossi Matias wrote in a blogpost about the technology.
During the demonstrations, the virtual assistant did not identify itself and instead appeared to deceive the human at the end of the line. However, in the blogpost, the company indicated that might change.
“It’s important to us that users and businesses have a good experience with this service, and transparency is a key part of that. We want to be clear about the intent of the call so businesses understand the context. We’ll be experimenting with the right approach over the coming months.”
Source: theguardian.com
