Steven's Blog

AI-Induced Atrophy

For some, artificial intelligence and automation in business herald a new era of increased productivity. For others, these are a harbinger of job obsolescence and layoffs.

I, however, question to what degree artificial intelligence and automation in the workplace will cause skills requiring human judgment and empathy to atrophy, and the impact that will have on effectiveness.

I have encountered a number of companies where artificial intelligence is already used to identify potential high-performing employees early on in their careers. Typically, I would expect a manager in the business to regularly assess staff for this purpose, whether as part of a formal process in the organization, an individual one, or both. I would not expect a human manager to defer to the judgment of a machine. Yet I find this is what is beginning to happen.

The HR director of a major company here in Japan told me once that she relies on an artificial intelligence system to identify the best candidates for employment and promotion, mistakenly believing a machine is unbiased compared to her or any other human. Amazon not long ago developed a system to identify the best managers with strong leadership capability in order to eliminate human bias. The system unexpectedly, however, displayed an aggressive bias toward men, never recommending women candidates. The machine merely mimicked male-dominated leadership biases of the company with a perfection even humans could not match. Amazon rightly abandoned the system.

Unbiased machines dont exist. Machine learning systems merely perfect and exacerbate the biases of whoever programmed them, or the bias present in human decision-making examples from which the machine learns.

At a recent event in Tokyo on artificial intelligence and automation, a panel of experts agreed that artificial intelligence is not intended to replace human judgment, but rather to improve upon it.

Just because a manager gets advice from a machine, does not mean he or she has to take it, only consider it. Click To Tweet

Yet what do you think the propensity of a human manager is to decide against the advice of a machine? After all, should a manager follow the machines advice and the advice turns out to be wrong, a manager can always blame the machine, whereas if the manager decides against the machine and the machine turns out to be right, the manager might be faulted for arrogantly ignoring the expert system. What if Amazon had not abandoned its skewed system?

Do you think it is far-fetched to presume that managers would abdicate decision-making authority against his or her better judgment to protect his or her career? Have you ever heard the expression, No one ever got fired for hiring IBM?Will no one get fired for deferring to a machine, or at least is that what managers will think?

At what point does it become so much easier to rely on technology by default that we forget how to do for ourselves what technology does for us? If you think such a scenario is unlikely, consider how many Japanese adults have lost the ability to write Chinese characters by hand because of their reliance on Japanese word processors? When it comes to automated driving, how long do you think it will be before people forget how to parallel park a car, if not how to drive themselves?

In a prescient 1966 episode of science fiction TV series Star Trek, the M5 advanced artificial intelligence computer ostensibly designed to command and run a starship without the need for a captain, officers and crew complement is being field tested, much to the dismay of starship captain character James T. Kirk, who is apprehensive about being replaced by a machine. Men no longer need to die in the icy cold of space!the M5s genius inventor proclaims.

The M5 performs brilliantly in tests until it takes a war games mock attack as a real threat and destroys a ship in its own fleet killing its entire crew. The M5 computer, for all its superlative intelligence,lacked the human judgment and empathy to tell the difference between a real and mock attack. Try as they might, the crew onboard the ship of which the M5 was in control was unable to turn the M5 off before it was too late. You cannot outsource human empathy and judgment to a machine. In the show, the M5 was abandoned.

Not so long ago, the autopilot of the most advanced Boeing 737 in service plunged the jet into the ocean killing all of its passengers and crew. A subsequent investigation has revealed that the autopilot incorrectly thought the aircraft to be in a stall, and responded by bringing the nose down. The pilots of the craft were unable to bring the nose back up as the autopilot continued to force the nose down again. Try as they might, the human pilots were unable to disengage the autopilot.

Had the pilots been able to disengage the autopilot, at least they would have known how to fly the plane. I shudder, however, to think that there may be a day when there will be no one on board who knows how to fly the ship without reliance on a computer.

And what about in your business? Who will still know how to fly your ship in the not so distant future?

I have identified five common characteristics of robust strategy among my most successful clients.

If you would like to know what these are and where your business stands, send me an email, and I will send you a free PDF.


Leave a Reply

Your email address will not be published. Required fields are marked *