All too often we hear startups pitching their product as having “algorithms” and therefore Artificial Intelligence or AI. The term AI and algorithms are tossed about like popcorn in a movie theatre. And algorithms are simply the butter you lavishly heap on your popcorn; makes it taste better, but it’s not AI. So what are algorithms and how do they apply to AI? You can’t have AI without algorithms, but they aren’t AI in and of itself.
An Algorithm Explained
Quite simply, an algorithm is nothing more than a recipe. It’s not the ingredients that go into the recipe, it’s the process, the steps in a recipe. An algorithm would tell you the process to make a stew through each step; cook stew meat, finely chop vegetables, add beef broth, add in broth and vegetables to meat, heat until ready. You can use the same algorithm every time you make the stew, except you may add different spices or vegetables, which is the “data” or ingredients. Humans have been using algorithms for thousands of years.
Getting to Artificial Intelligence
Algorithms are at the core of AI. Laid over top of algorithms are probability calculations. These calculations use data and today, often massive amounts of data, what we now call Big Data. Then there’s machine learning, natural language processing, neural networks, cognitive computing, pattern recognition and so on. AI can consist of these various elements mixed together in different ways. Algorithms are the processes of bringing all these ingredients together and telling them what to do.
This is a highly simplified explanation of what makes up AI and the role of the algorithm. But an algorithm alone is not AI.
The Rise of the Algorithm
As our physical world has become fused with the digital world (phygital if you will), algorithms play an ever more important role. It is algorithms that are the secret sauce for Google for their search engine. For IBM Watson and other advanced computing platforms. Their importance can’t be overstated. Creating really good algorithms is an exceptional skill that pays very well if you are good at it.
To a large degree, the word “algorithm” has been latched upon by marketers (I’m guilty of this) as well as startups and news media to the point it’s almost become a sort of meme. If you understand the place and role of an algorithm however, you can better understand how emerging technologies are being pitched and packaged.