Computers can only be as right as the people who write the codes and create and input the data.
* By Fiona Soltes *
Annette Danek, SVP Fulfillment at Penguin Random House, has had the opportunity to learn coding at several different points in her career. Each time, she said, she’s been struck by the amount of power at her fingertips.
“Do the wrong thing, and you can take down the system,” she said with a chuckle. “You can really make a system great, or you can make a system bad by coding it the wrong way.”
It begs an interesting question: Human error is always a possibility, but what about the machine? In our age of increasingly artificial intelligence, can a computer be wrong?
Dan Freeman of MHI member Evanhoe & Associates rose to the challenge of offering an answer.
“Of course,” he said. “Computers can only be as right as the people who define them or create them or create that data.” Computer algorithms will work as they have been told, Freeman said, “but if the algorithm is wrong, you’re always going to have a wrong answer… I guess, then, the true answer is that the computer is always right. It was the human that put the information there in the first place that was wrong.”
When it comes to artificial intelligence (AI), he said, it’s really about the amount and quality of data that’s available. Decisions are made based on historical data, and if there is not enough data available—or the data is not the correct kind—it won’t work.
But there’s something else at play, too: “AI is such a broad term, and people use it for everything,” he said.
“Things that are really just predictive analytics, people are considering AI, when there’s no machine learning whatsoever. It’s just a human coming up with some algorithms, and trying to predict the future based on the past. We’ve been doing that forever.”