Who will manage the artificial intelligence machines?

One of the largest, most successful hedge funds in the world, Bridgewater Associates, will be using artificial intelligence to remove the element human of emotion from decision making. In the process, humans will move from not only writing the code for these machines but also developing the boundary conditions for these decisions. In essence, humans will now program & manage machines. One could argue that humans have always managed the productivity of technology assets. We decide when to purchase or retire computers and servers. However, the key important distinction is that the next generation of machines are not only considered intelligent but in some aspects (emotionless) superior to humans for a given tasks or set of tasks.  While clearly human involvement is still necessary to program the machines, provide the boundary conditions, analyze the performance data and incorporate enhancements, it should be obvious that a much different set of skills will be required for the artificial intelligence manager of the future. She will in essence be asked to manage not only teams of people but also artificial intelligence assets in a way that drives value for the organization. So what skills will be required for success?

 

One of the key skills in successful people management is understanding human motivation and aligning incentives to drive target outcomes for the organization. Since there is no incentive, per se, that you can provide a computer like increased pay or larger bonus or more responsibilities, there will need to be an intuitive and detailed understanding of how artificial intelligence works. From the underlying theory to the practical applications to the key risks associated with the rapid advancement of computer capabilities. These will not be easy levers to manipulate. But, those who become skilled in improving the performance of artificial intelligence machines will be highly regarded and rewarded in organizations.

 

In modern day organizations, there is an annual budget developed toward the end of the year for objectives to be accomplished in the coming year. Increasingly, objectives will be geared toward which activities can be managed by an artificial intelligence machine. There will have to be clear work scopes and an assessment of available resources that are either off the shelf or that can be developed internally. This goes beyond the simplistic cost assessment to a more nuanced evaluation of can a smart machine increase the performance of the organization and how long will it take to accomplish. In many cases, the team of people who may be replaced will have to play a big role in having their work tasks deconstructed into units and watching a machine automate these tasks with increasing precision before either moving on to a new role within the organization or leaving entirely.

 

For most work streams, simply seeing a one time bump in performance will not be enough. Organizations who invest in artificial intelligence assets will expect to drive increased performance over multiple years. This will mean iterating upon an increasingly complicated algorithm that is simultaneously learning and improving automatically. Since the technological advancements or expectations will already be built in, the manager will have to take that even further in order to be considered successful. In essence, she will have to learn at a rate equal to or greater than the machine to deliver differential performance.

 

While it is somewhat predictable that a successful hedge fund would pursue an artificial intelligence strategy - they have the resources to invest early and iterate until they figure out the model, there is no doubt that as early adopters see success, more and more companies will pursue these options as well. And, as we see an increase in artificial intelligence machines, there will also be increased demand for those who have the skills to successfully manage these technology assets.