If there has been one takeaway from Satya Nadella’s recent keynotes, it’s that Microsoft is by no means falling behind in the race to produce workable artificial intelligence. In an emphatic style befitting of a major tech CEO, Nadella declared AI to be ‘at the intersection of [Microsoft’s] ambitions,’ before comparing its arrival with that of the book, or the internet. Such is the scope of application of the technology - which essentially helps computers independently recognize trends in avalanches of data - that no major technology company is failing to invest in it.
Arguably, no company is throwing itself into AI quite like Microsoft. The past few years have seen the company delve deeper into the b2b space, with its CRM tool Dynamics 365, in particular, driving a huge chunk of its revenue. And it too will be affected by the company’s commitment to AI. 'Take something like sales,' he said. 'In any business application you always explicitly modeled the world […] but there is one real problem: most of the sales activity happens outside of a CRM system. And so the goal of intelligence is to be able to reason about your sales data model. Not just inside your CRM system but outside.'
Some of Microsoft’s application of AI may seem gimmicky - stories like the company teaming up with Liebherr’s appliance division to put together a refrigerator that can identify the contents within itself using Microsoft Cognitive Services Computer Vision API, for example, will not change lives. But the extent to which machine learning will underpin Microsoft’s wider range of products is something Nadella has been keen to push. In a swipe at the likes of IBM and Google, the Microsoft CEO insisted: ‘We are not pursuing AI to beat humans at games,’ rather the company is looking to ‘democratize AI.’
Nadella is obsessed with the notion of empowering the user to create things, an MO that bleeds into its vision for Microsoft’s AI. The idea of ‘democratizing AI’ is, essentially, that the company wants to make it easier for software developers to build AI technologies based on Microsoft’s platforms rather than its competitors’. Then, it hopes that AI technology will become so widespread as to be mainstream.
One element of AI design that Nadella has been particularly focussed on has been ethics - particularly important when others are invited to create AI technology themselves. Quartz grandiosely referred to it as his ’10 commandments’ for AI design, but Nadella’s list of guidelines is difficult to criticize. Essentially, he believes that AI should only be designed to assist humanity, it should be transparent, it shouldn’t serve any particular group over another, it should protect privacy, and it should always have a kill switch. The company is still working out what it means ‘to have algorithmic accountability when you’re training a deep neural net,’ according to Nadella, but is using the problems raised by ethics as ‘guiding principles’ in AI’s development.
Nadella’s first book - Hit Refresh - will focus heavily on his belief that intelligent machines will reshape society, and he seems more optimistic about the near-future effect of AI than the likes of Apple, who have been largely quiet. The reason for Microsoft’s doubling down on artificial intelligence? It fears losing another major battle. Microsoft is playing catch up behind the likes of Google and Apple with regard to smartphones and personal computers, and it will be hoping it can reinstate its position as a leader when the wave of AI technology breaks. Nadella has been CEO of Microsoft for little over two and a half years, and has been successful in steering the company toward what he perceives to be its strengths. If AI becomes anywhere near as influential as he expects it to, he may have steered Microsoft toward dominance.