One of the key questions about the skills of the future is how humans differ from machines. And what are the fundamental differences? The big humans and machines see problems differently. The answer to this will determine the skills people should focus on developing, and the work that only people will be able to do in the future.

Machines, and here I mean computers and information technology in particular, need data. Machines like simple and well defined bits of information, so numbers, values, and figures are their favourite foods. If you tell a machine that a glass contains ten blueberries it will understand you. But if you tell a machine that a glass contains a lot of blueberries it will be confused. A machine would understand if you told it there were the square root of 100 blueberries, 60% of the volume of the glass of blueberries, or a random integer between 9 and 11 of blueberries. But so far it would have no ability to process words that require understanding context or complex interpretation.

Machines will continue to develop at tremendous speed, but ambiguous expressions will still be as difficult for machines as they are sometimes for humans. And what if I had used an example that involved the words good or evil. The machine would have been just as confused. Defining good and evil requires a human. We can argue whether blueberries are good or bad, or whether the glass is half full or half empty. We can tell why there are blueberries in the glass in the first place, and talk about how they look as a still life. We can even squabble about all of this, but a machine can only give us a numerical value. We are optimists and pessimists, but fortunately machines are neither.

Imagine an ant running across the floor and jumping onto your head. This idea is silly because we know that an ant can’t jump onto your head without superpowers. We’ve seen ant hills, and we’ve learned that ants, despite how strong they are, they can’t jump anywhere near that high. But what if I gave the same situation to a computer to evaluate. What tools would the machine use to solve this problem? Well, a machine can’t remember its childhood, or the summers it spent playing in the woods, or the first time it encountered an ant. It’s also never poked an ant hill with a stick. A machine has to know all the numbers because without them it can’t solve problems. So, how big is the ant? Or how long are its legs? What is the muscle mass? How strong is it? How does it jump? How tall is the target? The equation becomes complex, but that’s not an overwhelming challenge for the machine. In the end, the machine, maybe, might come up with an answer that an ant can maybe jump, let’s say, 0.0321795 centimetres, but not, let’s say, two metres. Right? So a person and a machine, they come to the same answer. An ant can’t jump that high. But they reached the conclusion through different kinds of thinking.

Tasks that require contextual understanding, situational awareness, and interpretation related to culture, history, or social norms are tasks where we humans are a bit superior to machines. We have all the information we’ve accumulated over our lives, and a machine only has the data that’s been given to it. Even if a machine has a million data points and has reviewed more information than any human alive, it will still give its answer as a probability that only a person can interpret as right or wrong, good or bad, a lot or a little.

The faster technology evolves, the more deeply we need to understand and interpret humanity. Ultimately, technology is only a reflection of us. It makes us strong, and it gives us influence, but it doesn’t change who we are. Technology can’t teach us what’s important, for example, and that is left to us because it requires interpretation.