Artificial Intelligence
Feb 14, 2018

Don’t underestimate importance of process in coming world of AI

​To win the game, your workforce must boost machine intelligence​

Much has changed in the 20 years since the famous Garry Kasparov versus Deep Blue battle for chess supremacy in 1997. Certainly, machine power has continued to advance according to Moore's law, software has evolved into cloud services, and AI has become an early stage reality. But has human understanding advanced regarding how to best use these new tools? More specifically, have we learned how to best exploit these tools within our work environment? So how can human traits, combined with industry-specific knowledge, and process, enable success?

Chess as guinea pig of early AI

AI researchers have long regarded chess as a vehicle to perform research. Before 1996, computers were no match for chess grand masters, but that rapidly changed after Deep Blue's victory over Garry Kasparov, who many experts consider to be the greatest grand master of all time. In the ensuing years, machines have advanced to the point where today, no human chess grand master could defeat a machine programmed to win at playing chess.

In more recent years, chess tournaments have taken place in which humans have teamed with computers and competed against other human-computer teams. One would suspect that a strong human + machine would beat a weak human + machine; however, surprisingly, this was not the case. The winners weren't humans who had the most powerful machine or some secret chess software. Nor did they let the machine make all the decisions. Rather, winners were often relatively amateur chess players who depended more on the interaction and coaching of their machines during the games. It seemed the process for human interactions with their machines was what made the difference between winning and losing on the chess board.

Garry Kasparov analyzed these human-machine chess competitions over the years and came up with a formulation that has been called “Kasperov's Law," which says: weak human + machine + better process is superior to strong human + machine + inferior process.

In addition, Kasparov has written that in chess competitions where teams are allowed any combination of people and/or computers, he found the teams of human + machine dominated even the strongest computers. Kasparov stated, “Human strategic guidance combined with the tactical acuity of a computer was overwhelming."

As I've written before about the advantages of human + machine, this teaming offers humans the ability to focus on higher value-added — and more enjoyable — tasks. Within the context of chess, Kasparov says it well: “When playing with the assistance of computers, we could concentrate on strategic planning instead of spending so much time on calculations. Human creativity was even more paramount under these conditions."

The way forward in world of AI

Let's move off the chess board and into the world of work. Companies around the world are experimenting with AI-based systems and recent Genpact Fortune Research indicates many leaders have not yet prepared their workforces for how to interact with these machines.

In addition, the World Economic Forum discusses the importance of reskilling the workforce in the machine age and states: “Digital technologies fundamentally transform organizations, with the pace of technological change exacerbating the challenge. Organizations must have a coherent strategy that includes a plan to reskill workers."

This reskilling must consist of helping employees understand how machines can be helpful to get their work done and how to best interact — and team — with them. This human + machine teaming will lead to breakthroughs across many industries and different business functions.

The key ingredient for successful AI deployments in the enterprise will be designing the processes we use for managing the machines. Only humans can do this — ideally, humans with knowledge of the particular domain and business process. This process knowledge and human ability to interact effectively with machines will determine success or failure in the coming AI world.

Bring in a trusted advisor or qualified partner if necessary, but don't delay. Experimenting with new AI-based solutions and investing in workforce reskilling needs to become the new normal. Otherwise, for some companies, it will be “checkmate."

About the author

Dan Glessner

Dan Glessner

Vice President, Digital

Follow Dan Glessner on LinkedIn