Artificial intelligence is expected to erupt over the next decade. But clearly, it’s still a “brittle beast.” Take this soccer game as an example.
A Scottish soccer team, hoping to cut down on the amount of staff needed to live stream games, used an automatic AI camera tracking system to ensure their audience could watch as the ball made its way around the field.
The problem? One of the linesmen is bald, which confused the AI:
As you can see, the AI thought that shiner of a scalp was actually a soccer ball.
According to a Verge report:
Pixellot, the company that makes the camera technology used by Inverness Caledonian Thistle, confirmed to The Verge that the problem was caused by visual similarities between the linesman’s head and the soccer ball. They noted that the angle of the camera didn’t help, as it made it seem as if the linesman’s head was inside the boundaries of the pitch, and the game ball itself was yellow, which added to the confusion. The company said the error was fixed shortly after the game ended.
I remember my eighth grade computer science teacher once told us: “a computer’s main problem is that it does exactly what you tell it to: nothing more, nothing less.”
At that time, prior to huge advancements in artificial intelligence and, truly, before the phrase “AI” was even a thing we heard much, I took this to mean that I needed to know all the ins and outs of my Dell desktop, else I risked entering a prompt which wouldn’t make sense to the machine.
But, of course, that teacher’s words have much more weight these days.
AI: it’s beautiful, and it’s coming to a project near you…but it’ll only work reliably if you know how to program for the unexpected. Otherwise, it’s just going to plod along, doing the thing it was taught to do, even if it can’t distinguish between something a human could with ease.