The phrase “artificial intelligence” still stirs up futuristic thoughts of the “terminator movie” kind of AI. AI that resembles human intelligence. The reality is that AI is already all around us, especially in the business world. Businesses that learn to utilize AI can give themselves a significant advantage over those that don’t.


Think of AI less as a human brain that’s very smart generally and more like a singularly focused machine. Human smarts are different from computer smarts in that we tend to know about a lot of topics and can draw abstract associations between them. You can’t expect a computer to be creative in the traditional human sense, but you can train it to solve specific problems. This is where AI comes in. How does this manifest?


1. Machine Learning


Machine learning involves feeding a computer lots of information about a specific problem so that it learns how to solve it. For example, if you’re a real estate appraisal company, you might have the particular issue of “being able to give accurate appraisals for houses in your area.” How could you teach a computer to conduct accurate appraisals? You feed it information or data. 


The data in this example would take the shape of information about past real estate sales. You’d give the computer hundreds (or even thousands) of examples of how much money a given house sold for and the different characteristics of the house (aka the number of bedrooms, presence of a pool, distance from a school, etc.). Over the course of all the examples, the computer will learn more and more about how much each characteristic of the house contributes to the overall cost.


Finally, when a new house comes along without a price tag, the computer uses all of its past learnings to make a prediction about the price of the house (or conducts an appraisal). Voila, narrow problem solved!


2. Deep Learning


Deep learning is the next level up from machine learning. Deep learning methods more closely resemble the human brain and can thus be used to solve more generic problems, like predicting the next word in a sentence. The house appraisal problem is pretty narrow. However, predicting the next word in a sentence is highly general! I could be writing about anything right now; sports, music, TV, science, technology. Predicting the next word in any sentence is a far more general problem because the sentence can be about anything.


With more general problem solving comes more advanced AI.