With all the artificial intelligence propaganda that’s been showing up these days, discerning marketing fluff from truly intelligent solutions isn’t easy. Discerning the difference could save you from investing in a dead-end solution.
And if you’re secretly wondering what A.I. really means; you’re not alone. The problem is not figuring out whether something is artificially intelligent or not. The problem is that A.I. doesn’t really have a proper definition to begin with.
In fact, there isn’t even a single definition for just the word “intelligence.”
But there’s at least 70 in case you want to pick your favorite. So, if we can’t make up our minds on what intelligence is, we certainly won’t ever achieve consensus on what artificial intelligence is.
Now that we’re clear we can’t ever be 100% sure of what A.I. should really be, we can at least say what it should or shouldn’t look like, what it should do for us, and how it should behave.
Most A.I. propaganda shows images of robots and machines that resemble a human. And that makes sense – if anything could be called intelligent, it’s got to be a human, right?
Zombies also resemble humans. And thankfully, they’re far from any definition of intelligence. (So, do we really have to fear a zombie apocalypse? What about A.I.-powered zombie humans?)
The real thing to worry about as you're considering spending precious real dollars to buy into A.I. is that you’ll only get a zombie in a robot costume.
So, how do you know that an A.I. solution really isn’t a zombie?
How to Spot a Zombie Dressed as a Robot
A zombie is kind of like a robot. It’s self-sufficient. It doesn’t need a lot of guidance and direction to do the things it does. But it’s stubbornly un-trainable. You can’t get a zombie to do anything it doesn't want to (and you'll never know what it's thinking!).
The easiest way to spot the zombie inside a robot is to open the robot up. Robots contain no mysteries. They’re fully documented and they produce consistent results based on specific settings, controls, and models. They even have known limitations.
Zombies are A.I. solutions powered by black boxes. The problem with black box "A.I." is that there is no accountability for how it works. Most black box solutions are really just a bunch of API connections to other services, or worse - algorithms with undisclosed training.
But a robot is well thought out, mechanical, and reliable. There are no mysteries, no hidden APIs, and require no “magical” tuning to get accurate results.
The A.I. that will go to work for you is the same. Demand explainable results, and visibility into the algorithms that make it work.
Here's an example of how transparency is built into an A.I. automated intelligent document processing software:
