Microsoft buys conversational AI startup Semantic Machines to improve Cortana
Microsoft has been pooling a lot of resources into artificial intelligence recently.
Take Bing, for example. While often ditched in favour of Google, it is continuously being made more intelligent and nuanced.
Now the company’s set to make another core aspect of the Microsoft experience more intelligent as it acquires AI startup Semantic Machines.
Semantic Machines is a California-based startup that develops conversational AI capable of speaking to humans in a way that sounds natural and that allows conversations to continue past simple queries such as: “What’s the weather like?”
It’s an acquisition that has been made in an effort to improve the Cortana voice assistant as well as Microsoft’s social chatbot XiaoIce, which has had more than 30 billion conversations on Chinese social networks since it was first released in China back in 2014.
“With the acquisition of Semantic Machines, we will establish a conversational AI centre of excellence in Berkeley to push forward the boundaries of what is possible in language interfaces,” Microsoft wrote in a blog post. “Combining Semantic Machines technology with Microsoft’s own AI advances, we aim to deliver powerful, natural and more productive user experiences that will take conversational computing to a new level.”
Semantic Machines has some pretty talented people at the helm. It’s being pioneered by tech entrepreneur, Dan Roth, as well as two prominent natural language AI researchers: Dan Klein from the University of California, Berkeley and Percy Liang from Stanford University.
While Microsoft highlights that it made the first “full duplex voice sense” chatbot system – a term that refers to the ability to communicate with an AI in both directions simultaneously – in XiaoIce.
Google recently made the same breakthrough in phone calls. Earlier this month, Google announced at I/O 2018 a duplex AI system that can make phone calls to shops and restaurants on behalf of a human. It hasn’t been without controversy, however, with concerns being raised about the possibility for duplicitous impersonations and further intrusive Google data mining.