At the end of last month, Sundar Pichai wrote his first annual shareholder letter as CEO of Google. Eight months prior, Larry Page and Sergey Brin led a restructure of the company they founded, separating its core internet business from subsidiaries focusing on new areas of innovation, such as self-driving cars, drones, augmented reality, biotech and life sciences. Page and Brin now lead umbrella company Alphabet as CEO and president respectively, while Pichai heads up Google.
It was the first time the annual shareholder letter had been written by anyone other than Page and Brin. Pichai kept it simple, outlining the key areas Google will focus on across its product lines – but he was bold enough to say they will all be driven by a long-term investment in machine learning and artificial intelligence.
Pichai wrote, ‘It's what allows you to use your voice to search for information, to translate the web from one language to another, to filter the spam from your inbox, to search for "hugs" in your photos and actually pull up pictures of people hugging… to solve many of the problems we encounter in daily life. It's what has allowed us to build products that get better over time, making them increasingly useful and helpful.’
Google has been building the best AI team and tools for years, said Pichai – a claim he backed up by recalling the breakthrough its AI subsidiary DeepMind made in March when it became the first computer in history to defeat a top-ranked human player of the ancient board game Go.
‘This is another important step toward creating artificial intelligence that can help us in everything from accomplishing our daily tasks and travels, to eventually tackling even bigger challenges like climate change and cancer diagnosis,’ said Pichai.
Google is betting on machine intelligence as one of the most important innovations to disrupt the world over the next ten years and beyond. And it’s not alone, competing vigorously with the likes of Facebook, Apple, Amazon, Microsoft and IBM to be a leader in this transformative technology when its applications start translating to big bucks. DeepMind’s victory peeved off Facebook’s AI chief so much that he claimed it is 'completely, utterly, ridiculously wrong' to think it was 'true artificial intelligence'.
For that reason, Google is not throwing all of its eggs in one basket. DeepMind was acquired in 2014 and is working on various applications of machine learning but with a particular focus on healthcare. Earlier this year it partnered with the NHS to develop a smartphone app that alerts doctors and nurses of patients at risk of kidney failure. But DeepMind is not the only part of Google using machine intelligence.
‘There are currently over 100 projects that we are applying machine intelligence to,’ said Thomas Davies, director of Google for Work in Northern, Eastern and Central Europe, addressing the AI Summit last week. ‘These include search algorithms, driverless cars and medical diagnostics through Calico, one of our Alphabet companies.
‘Quite frankly, because it’s open-ended, it’s endless. We have a lot of products and services across the entire Google and Alphabet brands that we can apply it to. We are incredibly excited and within the enterprise space you’re going to see more and more of these machine learning applications coming through from Google and our partners.’
As well as developing its own applications, part of Google’s machine learning vision is to open-source some of its APIs. A distributed version of machine learning software TensorFlow, which powers Google applications such as its translations services and photo analytics, was opened up for free download last November.
By doing this Google is recognising the vast opportunities that machine learning can tap into, only several of which it can really focus on through its own (albeit huge) resources. Despite only being released in November, TensorFlow became the most forked project on code repository GitHub for the whole of 2015.
GitHub is somewhat of a bible for developers, so being the most popular machine learning framework on the site goes a long way to cementing Google as the leading authority in this field amongst such a vital part of the technology community.
‘We’re going to continue to build what we think are high performing, really useful applications using machine intelligence,’ Davies told delegates at the AI Summit. ‘That’s super important, but it’s not far enough, so we actually decided to open-source TensorFlow to the community.
‘I bet if about 20 of us got together and decided to build from the bottom up, from scratch, a new form of application to serve something, the value is exponential. That’s why we’ve decided to open-source: so enterprises can just go build applications themselves. That becomes incredibly exciting for the CEOs, CTOs and CIOs we work with.’