Microsoft Doubles Down on AI with 'Conversational Computing'
- By Jeffrey Schwartz
- March 30, 2016
At Microsoft's Build conference on Wednesday, CEO Satya Nadella laid out his company's vision for the future of artificial intelligence (AI) and machine learning.
Microsoft has already made plain its ambitions to make Cortana and its machine learning-based engine more useful and pervasive. Nadella elaborated on those ambitions during Wednesday's Build keynote, describing Microsoft's concept of "conversational computing" and how it will be integrated in everything from Windows, its Edge browser, Outlook and Skype, to the entire ecosystem of software, hardware and cloud-based SaaS offerings. Nadella sees conversational computing using speech as the next user interface and framework for how all applications are developed and connected.
In a callback to the name of the conference, Nadella called on developers to "build" the next wave of applications based on tools and services that Microsoft is rolling out, including 22 Cognitive Services APIs for the Azure-based Cortana Intelligence Suite (previously called the Cortana Analytics Suite) and the new Microsoft Bot Framework. Both the Cognitive Services APIs and the Bot Framework are available in preview, Microsoft announced Wednesday.
The Bot Framework is designed for developers building intelligent bots that use natural language across platforms and apps. Nadella said the Bot Framework, which includes an intelligence runtime, will work in any programming language and will allow developers to build interconnected applications that use speech as the UI.
"We think this intelligence runtime [in the Bot Framework] and Cortana Intelligence Suite is going to be core, just like how the .NET runtimes were, to all of the applications you are going to build," Nadella said. The notion of the Bot Framework runtime, Nadella said is that developers can build intelligence into their apps and ultimately use voice as the new UI.
Microsoft will roll out the framework over time, though the company is bringing it to Skype with the Skype Bot Platform, which consists of an SDK, API and workflows available in the new Skype Bot Portal.
"We want to take that power of human language and apply it more pervasively to all of the computing interface and the computing interactions," Nadella said. "To do that, you have to infuse into the computers and computing around us intelligence. It means you have to bring forth these technologies of artificial intelligence and machine learning so that we can teach computers to learn human language, have conversational understanding, teach them about the broad context...so that they can really help you with your everyday tasks both at work and elsewhere. So as we infuse intelligence into everything. I think it's very important to have a principled approach to guide our design, as well as how we build things."
Microsoft is not alone in its push to make computing and mobile devices that interact with users through speech. Apple, Google and even Amazon with its Echo device have similar offerings, though they are generally considered novelties due to their frequently unpredictable performance.
Microsoft's AI efforts have not always gone smoothly, either, as Nadella acknowledged during Wednesday's keynote. The company suffered a black eye last week when its new social media "chat bot" called Tay began Tweeting racist, anti-Semitic and sexist remarks. Microsoft disabled the bot and apologized.
"We quickly learned it was not up to this mark, and so we're back to the drawing board," Nadella said.
Jeffrey Schwartz is editor of Redmond magazine and also covers cloud computing for Virtualization Review's Cloud Report. In addition, he writes the Channeling the Cloud column for Redmond Channel Partner. Follow him on Twitter @JeffreySchwartz.