Microsoft Says Its AI Bot Can Also Make Google Duplex-Like Calls

Xiaoice, translated to "little Bing" in Chinese, is the name of Microsoft's ambitious AI-based chatbot that the tech giant has been testing for a while in China. The company, in a blog post on Monday, announced new features for this relatively unknown chatbot. These include enabling "full duplex" capabilities for Xiaoice for partners and developers and the ability to create customised audio stories in a short period of time. This free-of-cost audio service will be available for users in Asia starting June 1. Microsoft also announced that it will be acquiring conversational AI company Semantic Machines.

In the blog post, Harry Shum, Executive Vice President, Artificial Intelligence and Research, Microsoft explained Xiaoice has had human-like "full duplex" conversational skills (i.e. ability to have "human-like verbal conversations") since last August and that the AI software has talked to over 600,000 users on WeChat in the past few months. However, unlike Google Duplex showcased at I/O 2018, Xiaoice can talk only to the owner instead of a third-party business as of now.

Taking in consideration the allegations against Google's "unethical" demo of Duplex, Microsoft claims that the company has ensured fair, transparent, and accountable machine learning that includes letting users know that they are talking to a chatbot and not an actual person.

As for audio stories, Xiaoice uses input from parents and children to create 10-minute long personalised audio stories within a span of 20 seconds.

Microsoft operates three other chatbots - namely, Ruuh in India, Rinna in Japan and Indonesia, and Zo in the US. Zo is successor to the infamous Tay chatbot that was shut down within days of its release citing racist tendencies. We are yet to see any advanced conversational features on Microsoft's Cortana virtual assistant.

"As we develop AI, it's our responsibility to raise questions and study the issues now so that our technology creates benefit and opportunity for everyone. We have to build AI systems that hear all voices and recognize all faces equally across our diverse world, and ensure that the tremendous power of AI is accompanied by transparency and accountability that puts human beings in the driver's seat. There is great possibility, but we have a lot more work to do," said Harry Shum in the blog post.