It seems everything is being called AI or ML nowadays, but these two words have a very specific meaning. In this video, learn how to hear about an application and correctly identify if it really is AI or ML, and if so, which one.
- What comes to mind when you hear the word robot? What do you first think of? What about artificial intelligence? While I don't know exactly what you see in your mind's eye, I'm certain you imagined something, and likely something rather elaborate, and maybe even just a little intimidating. People have been obsessed with robots and other forms of artificial intelligence for hundreds of years. At the time of the American Revolution, Europe was fascinated by a chess playing robot that was dressed up like the popular conception of the Turk. Spoiler alert, not a real robot, there was a small chess player hidden inside. The early movie Metropolis had a machine person who rose up against her creators, and how 9000 from 2001: A Space Odyssey reminds us why it's never a good idea to get on the bad side of a robot in deep space. Now seemingly out of nowhere, robot and artificial intelligence technology has become real and part of our lives. IBM Watson wins Jeopardy, beating the most successful human player at the time. Deep minds AlphaGo beats the world's best Go player in a game so complex, experts say it makes chess look like tic tac toe. Almost every website you visit recommends books, movies, and other products based on your ratings and habits. Siri, Google and Alexa answer our spoken questions. Where I live in Silicon Valley, I often drive alongside Waymo self driving cars. Boston Dynamics robots navigate complex spaces while being tormented by humans. Less visible algorithms can also affect our lives. Credit card companies and merchants spend over six billion dollars per year to detect and prevent credit card fraud. And one of the primary tools they use is a technology called Neural Networks. Artificial Intelligence technology makes lending decisions. Which may determine whether you get to buy your dream house or car. In social media algorithms decide which cat videos you'll find adorable, showing you more of them in order to keep you online longer. And in an application with much greater personal impact. Machine Learning can even recommend the length of convicts prison sentences to a judge. In casual conversation, we call all of this technology artificial intelligence AI, but in fact, there are two different types of tasks being done. Artificial intelligence and machine learning or ML. Machine learning is about analyzing a set of data based on a model trained on many other similar data sets. ML algorithms can classify, determining which numbers or letters I just wrote on that check. My sixes and my Gs are pretty hard to tell apart. Cluster. Figuring out different topical buckets of questions that your customer service organization is being asked. Recommend. Suggesting products or experiences you'll like, based on what it's seen and heard from you and others in the past. Triage. Looking at the thousands upon thousands of security alerts raised on an enterprise computer network, and deciding which ones we should investigate immediately, which ones we should review when time allows, and which we can safely ignore. Predict the future. Extrapolating current box office receipts for different types of movies, to predict what this summer's blockbuster hit will be. All of these machine learning algorithms are analyzing data and even recommending action, but they're not themselves taking action. That's left up to the human. The customer service executive, the consumer, the security operations professional, or the Hollywood investor. Machine learning just provides some really smart analytics. With artificial intelligence, we take things one step further, and actually put the machine in charge. Some cars today will complain at us, if we're approaching an obstacle or leaving our lane. That's machine learning, because we decide what to do about it. But self driving cars are AI. There's no person in the loop, they just steer, break and hit the gas. When Watson played Jeopardy, it just played all by itself. That's AI. Now in customer service, most applications that people casually refer to as AI are actually ML. The final decision or action is left in the agent or the customers hands. So let's call it by its proper name, machine learning. We're in a golden age of AI and ML. And these technologies are set to revolutionize how organizations deliver customer service. And remember, if the technology you're considering provides insights to the data, that a human can take action on, it's ML. If the system uses ML to take action without consulting a human that's true AI.