It ain’t rocket packs, not quite yet, but it does feel like the future. The smartest machines in history: self-driving cars, robots, bots that answer questions like humans. The smartest brands in history: Google anticipating your search bar entries, Facebook putting together a personalized news feed, Amazon knowing every embarrassing detail of your interest in cookbooks and novelty toilet paper, and Netflix inventing a whole new genre based on your viewing preferences (“Nostalgic Hawaiian Sci-Fi Family Cop Documentaries”).
All of these technologies are powered by artificial intelligence: systems based on invisible, complex systems that process more data than has ever been produced — learning from it, responding, and predicting.
It certainly wasn’t a straight shot to here. Not exactly. As a New York Times story reports, interest in AI in the 1980s created a minor boom. Without any viable business uses, though, it led to an “AI winter.” Now, more data and more processing power have made AI possible — for real this time. Consequently, AI investment is booming. Per the Times:
Funding in A.I. start-ups has increased more than fourfold to $681 million in 2015, from $145 million in 2011, according to the market research firm CB Insights. The firm estimates that new investments will reach $1.2 billion this year, up 76 percent from last year.
Now there are uses for this technology (beyond just “Gripping Violent Animated Pet Dramedies”). That’s because there are users: internet and mobile makes billions of connected devices capable of benefiting from AI.
Here are four ways AI will change everything:
Have you noticed how computers have gotten smaller while getting smarter? They’ve also gotten cheaper: now there’s a computer inside anything with an on/off switch. All of these newly intelligent devices — toasters to toothbrushes, thermostats and lightbulbs and cars — are now being networked, talking to each other, and businesses, and consumers. Why shouldn’t your car tell your house that you’re nearly home so that the house can tell the oven to preheat to the proper temperature for that fish it knows you just bought, because you made the purchase with your phone and your phone told it so? So, behind every device is a customer, and the next generation of customers expect a connected, smart experience. We’re talking about a lot of connected things: six billion of them that, says Gartner, will be requesting support by 2018. Those billions of connected things mean huge volumes of customer data. Businesses need to be smart about the way they gather, digest, and apply that data, which is the lifeblood of IoT … if it can be properly used.
A huge gap is growing between companies and customers. For all the data customers are creating, less than 1% is analyzed, such that 77% of customers say they are not engaged with businesses. There are so many ways to read data and so many conclusions to be drawn about customer behavior and preferences — yet most of this potential insight is falling by the wayside because businesses aren’t prioritizing the analysis of that data. New tools reveal useful insights about the customer. These insights exist along a spectrum of Intelligence: the most basic tools require you to “pull” information out of them, while the most intelligent tools “push” information to you, anticipating what you’re going to want to know. For the latter, we turn to machine learning.
With machine learning, computer systems can take all this customer data and build on it, operating not just on what’s been programmed but also adapting to changes. Algorithms adapt to data, developing behaviors not programmed in advance. Learning to read and recognize context means a digital assistant could scan emails and extract what it knows you’ll want to know. Inherent in this learning is the ability to make predictions about future behavior, to know the customer more intimately and not just be responsive, but proactive.
Big data and analysis produce patterns, and when smarter machines are able to read patterns and learn from them, they can figure out what might come next, make deductions that are better than just assumptions, and make conclusions that are better than just guesses. The promise of a “digital assistant” is not the robot voice that answers our questions about temperature and movie times, but that knows our patterns, learns from them, and reminds us to leave now in order to beat our record of arriving 2 minutes late 80% of the time. The system needs to be fed, and from that comes smarter machines, connected devices, and the ability to predict our needs and wants. The more quality information we can give to the system, the smarter it’ll get.
Next time, we’ll get into how machine learning actually … learns. Netflix didn’t come up with your interest in “Silly Nordic Sorcery Cartoons” out of nowhere, after all.