Healthcare consultant Jeff Gorke’s recent blog offers an interesting perspective on the current state of machine learning and AI and what we can expect in the near term.
As Mr. Gorke notes, it is important to set realistic expectations with respect to these two burgeoning technologies. It is even more important to understand the requirements for a successful AI implementation (beyond being a “crazy smart” programmer!).
The first step in solving a problem with AI is to clearly and concisely define the problem to be solved. Step 2 is to gather a plethora of empirical data regarding the outcomes of various inputs at each decision point in the task workflow under consideration. If the problem is well-defined and the inputs are expertly applied to produce a sufficiently narrow range of desirable outcomes, then AI can produce a solution with a high-probability of success in an amazingly short time frame.
But even the smartest AI can’t feed itself. Not at first. Every AI or machine learning program needs reliable accurate data and a narrowly defined problem in order to work its “magic.” But with the proper input, there is no questions that AI is a “game changer.” And it isn’t as far away as you might think.
It goes without saying that this is an oversimplification of the challenges and requirements associated with artificial intelligence. But AI and machine learning are here to stay and it won’t be long before they are both being utilized by companies around the world to gain a competitive advantage. As Mr. Smith said to Neo on the subway platform, the “sound” you hear from the AI community “…is the sound of inevitability.”
