This is the second article in my ongoing partnership with Noodle.ai to spread the word about the possibilities enabled by ‘good AI.’ My first article, which you can read here, explored my idea of ‘augmented intelligence:’ that machine learning augments human intelligence, instead of replacing it, to create a sum greater than the whole of its parts.

A Formula for Human-Machine Partnerships

In my last post, I described the revolution in artificial intelligence that algorithmic chess spurred. In 1985, I had crushed 32 computers in simultaneous play. In fact, given the state of computer chess at the time, I was most concerned that, had I lost even a single match, some might accuse me of throwing the game for promotion’s sake. But just 12 short years later, IBM DeepBlue ended the reign of human hegemony in chess.

Yet, despite the human triumph that DeepBlue was––for humans had created the machine––it did not live up to the long-held expectations of AI aficionados. Instead of playing like a perfect human, IBM played like a powerful machine, equipped with the brute force ability to look deeply into a position, but not the creativity of a human player. While others worried that AI chess would “perfect” the game, I began to suspect that, just as we humans need the power of machine intelligence, machines also need humans to augment their abilities.

In 2005, my theory was put to the test when the online chess platform Playchess.com hosted a “freestyle” chess tournament. The rules were wide-open––teams could employ as many human beings and as much computer help as they wanted––so it became a perfect experiment to test human and machine intelligence in collaboration. With several groups of strong grandmasters and powerful chess machines like Hydra in the mix, the competition seemed to be a test of grandmasters against chess machines. To everyone’s shock, the winners turned out to be a pair of American amateurs using three average computers, linked in a chain processing system to create a home-made supercomputer. The pair won by using human intuition to steer the reins of the raw, computing horsepower of machines.

Reflecting afterward on the tournament’s surprise ending, I first formulated what others later dubbed “Kasparov’s Law:”

  1. Weak human + Machine + Better Process > Strong Machine, AND;
  2. Weak human + Machine + Better Process > Strong Human + Machine + Inferior Process

The point is that process mattered most, not the raw brawn of machines or the artistic ability of humans. Process is the mechanism to target and structure which problems machines should solve and the selection of which recommendations to make to the human. The team that could best employ the strengths of humans and machines to augment each other was the team that won.

 

Progress Over Perfection

When AI first turned its sights on chess, commentators speculated an end to chess. Machines would achieve perfection, and there would be no point in playing. A real-life HAL 9000 might simply announce move 1.34, with checkmate in, say, 38,484 moves. Of course, those peanut-gallery Cassandras were wrong, and chess never will be ‘perfected.’ The game is simply too complex, with too many permutations and possible situations.

But the fundamental error was in chasing perfection and moonshots instead of the achievable markers of technological progress that have always brought human beings out of darkness and into the light, one small step at a time. Implementing augmented intelligence into our day-to-day life does not need to be revolutionary, and it does not need to solve all of our problems at once to still be progress. But by taking the new tools and lessons of augmented intelligence, we can gradually improve every aspect of our lives: from eliminating supply chain waste to automating away some of life’s annoying, thankless chores.

 

How Augmented Intelligence Can Bring Life to a Zombie Economy

But what excites me the most about augmented intelligence is its possibility to truly unleash the creative spirit that makes us all human. A 2015 McKinsey Co. report found that  just four percent of work tasks in US jobs currently require creativity “at a median human level of performance.” That means that 96 percent of the things US employees are asked to do is just zombie work. Those dull, banal tasks are already dead; they just don’t know it.

My friend, Noodle CEO Steve Pratt, works every day to eliminate the thousands of minute inefficiencies that add up to trillions of dollars in manufacturing waste every year. But what about the waste caused by forcing 96 percent of American workers to imitate machine labor? And, let’s be honest, the machines can do much of this work more efficiently than we can.

By applying augmented intelligence to our most boring tasks, we harness the computing power of machine learning and free up humans to do the creative labor at which our species excels. The end goal is to re-center labor on the human experience: making our jobs more enjoyable and each of us more productive. Instead of spreading stories about the potential dystopias of an AI future, we should seize this chance to bring life to a zombie economy.

In last year’s Earth Day conversation with Noodle, Steve stated that, given the new opportunities augmented intelligence offers us, waste is now a choice. Let me go one step further. When we choose to ignore the possibilities AI offers us, we are not only choosing wasted assets and wasted dollars, we are choosing to waste the time and creative potential of our fellow humans. And if we can just get over the one-inch tall barrier that is our fear and apprehension, we will find a new world of opportunities for enriching, fulfilling labor available for each of us.