The tech sector is currently "worshipping" a handful of artificial intelligence companies, but is yet to show how it will make the technology accessible to and useful for its human end users, Microsoft head Satya Nadella has said.
Speaking at the DLD technology conference in Munich, he said: "The way at least I have defined AI in simple terms is we are trying to teach machines to learn so they can do things that humans do, that in turn helps humans. Which is augmenting what we have."
He added: "In the next phase, to me, [is] how can we democratise this access, versus worshipping the four, five, six companies that have a lot of AI. To actually saying: AI is everywhere in all the companies we work with, in every interface, [and that] every human computer interaction is AI-powered."
Nadella called for AI that helps build on human capability and creates more trust in technology from a security and privacy point of view, as well as having "algorithmic accountability".
Microsoft is currently building support for AI into tools many developers can use. In 2016 the company released Cognitive Toolkit for developers interested in natural language processing on GitHub, and showcased Zo, its latest chatbot.
Last week it also attempted to advance its efforts in conversational AI by acquiring machine learning startup Maluuba to “solve artificial general intelligence” through language.
But when it comes to collaboration over AI, some companies are more open than others. For example, research from Google and Facebook tends to be public, while Apple only recently overturned its culture of secrecy and published its first official research paper last month. ®