The evolution of AI Assistants

Roberto Pieraccini, Director of Engineering, Google, Zurich (CH)

Session I Keynote: March 30^th, 09:15-09:40

The concept of AI Assistant, often referred to as virtual assistant, is not new in the field of human-machine natural interaction, and we can trace that back to the first experimental speech recognition systems built in the 1970s. As of 1987 Apple promoted a vision clip, called the Knowledge Navigator, that illustrated what a virtual assistant of the future may be like. It was not until 2011 that Apple exposed the millions of users of their flagship product, the iPhone, to Siri, the first consumer virtual assistant. Since then, we have seen Amazon’s Alexa, Micorsoft Cortana, Samsung’s Bixby, and The Google Assistant reach hundreds of millions of users around the world, in dozens of different languages, and via a multitude of different devices. But what is an AI Assistant? How does it work? Where is the complexity? What is the future vision? I will answer these questions using the example of the Google Assistant, and show how the evolution of machine learning is playing a fundamental role in the evolution, sophistication and intelligence of these systems.