We’re going to talk to our technology, and everyone else’s too. Google proved that earlier this month with a demonstration of artificial intelligence that can hop on the phone to book a restaurant reservation or appointment at the hair salon.
Now it’s just a matter of who can build that technology fastest. To reach that goal, Microsoft has acquired conversational AI startup Semantic Machines for an undisclosed amount. Founded in 2014, the startup’s goal was to build AI that can converse with humans through speech or text, with the ability to be trained to converse on any language or subject.
Microsoft is also acquiring some key talent in the deal: Semantic Machines’ CTO Larry Gillick, who was previously Apple’s chief speech scientist for Siri; as well as AI professors Dan Klein and Percy Liang, from UC Berkeley and Stanford, respectively.
While Microsoft’s conversational chatbot Tay was a disaster in the United States, an extremely similar chatbot called Xiaolce has been a resounding success outside of the US, racking up 30 billion conversations with more than 200 million users worldwide. Microsoft’s virtual personal assistant, Cortana, also lives in every Windows 10 computer, though it’s not as useful in a voice-only setting like a smart speaker.