Use Alexa, Siri, or Google Home? They are AI’s – Artificial Intelligences. Today, AI is in the press everywhere. But, only a short time ago it was a pariah. Cloud changed all that.
You can’t avoid them. Elon Musk, Bill Gates and Stephen Hawking forecast a dark future of mankind with AI transcendent: a world of Terminators or the Matrix. Others paint futures where they seep into your every device and make life more convenient. They are learning to read, speak and see; and they play Chess, Go and Poker better than humans. They seem inevitable.
Hard to believe but it was not that long ago the concept of AI was so pooh-poohed that there were no funds for it and researchers would not use the term for fear of ridicule. During the 1980’s and 90’s sentiment against the concept became so bad it virtually disappeared from any discourse. This was the AI Winter.
Why such an aversion? Call it the end of a hype cycle. After many years of work and billions of dollars spent there seemed to be little to show for it. Disillusionment amongst the researchers, the press and especially investors set in and the funds dried up.
Like most of these kinds of cycles in developing new technology, some work continued under different labels at a slow simmer but what was needed only emerged with the turn of the millennium. As the underlying algorithms and approaches were refined what was missing was vast quantities of data for an AI to learn from and inexpensive gobs of computing horsepower to process the data and the algorithms in a timely fashion.
Enter Google. Search queries provided astounding amounts of data and Google’s massive cloud data centers stored and manipulated it quickly and for a fraction of the cost of the way data centers were then being run. Google’s founders knew exactly where they were going. In 2000, Larry Page said: “Artificial intelligence would be the ultimate version of Google.”
Maintaining the pace Google has continued to build its AI expertise. In 2014, it bought Deep Mind Technologies who recently built the AI, AlphaGo, that bested the world champion in the ancient game of Go. And in a gesture to spread AI even more widely and help developers get a handle on machine learning, Google has open-sourced some of its platforms, including TensorFlow, which is an open source library for machine intelligence.
But until recently Google kept its computing horsepower and algorithms to itself. Where did other researchers get their hands on those kinds of resources? Enter AWS (Amazon Web Services) in 2007. Now, anyone could secure those needed “gobs” of computing power cheaply and for only as long as needed. AI researchers could experiment with all sorts of approaches and extend their knowledge.
Eventually, in 2016 AWS would launch AI as commercial services that anyone could use. The three services being rolled out are Amazon Rekognition for image recognition; Amazon Polly for text-to-speech services; and Amazon Lex, the technology inside its smart device Alexa, offering speech recognition services.
IBM was busy as well with its Watson. In 2011, it competed against two human former champions on Jeopardy and won, setting off a small firestorm in press coverage. Always a little behind, back then IBM ran Watson on a set of servers that could fill up the size of a master bedroom. Today, it runs on the cloud.
Just look around and you’ll see every product and service testing or touting its AI capabilities. In addition to the IBM’s, Amazon’s and Google’s, retail is bubbling with AI fever from Starbucks and Lowe’s to the Italian lingerie maker Cosabella.
The AI Winter is now well and truly over and while the new digital assistants, game playing programs and other algorithms are impressive, we are only in the AI Spring. In other words: Buckle up! And, don’t worry. We are a long way from Ken Jennings’ welcoming our computer overlords.