Google Duplex: An AI System for Accomplishing Real-World Tasks Over the Phone
Google Duplex is an advanced AI system that can hold natural conversations with real people over the phone to accomplish real-world tasks. With this technology, Google aims to make the conversational experience as natural as possible, enabling people to talk to computers like they would to another person.
Duplex is capable of generating natural speech, thanks to the application of deep neural networks, and can perform specific tasks, such as scheduling appointments, conducting restaurant reservations, or ordering food delivery, among others. However, to ensure that the conversation stays natural, Duplex is limited to closed domains that are narrow enough to explore extensively. It cannot engage in free-flowing general conversations.
The technology is built to sound natural, comfortable, and transparent. The voice is designed to sound like a real human, using speech disfluencies like “um” and “uh” and natural language rhythms to make the conversation as authentic as possible. Additionally, Google wants to be clear about the intent of the call so businesses understand the context.
Google intends to continue experimenting with the right approach over the coming months and refine the technology to make it a seamless part of people's lives.