The Unintended Consequences Of Machines Taking Our Simple Jobs

In the grand debate around artificial intelligence and how it will impact the workplace, one of the more persuasive arguments has been that AI will take on many of our more routine tasks, which will free up our time and energy for more complex tasks.

As a narrative it makes a lot of sense, as it matches the current capabilities of the technology in relation to the more human aspects of what we do.  This approach to man/machine interactions is sensible so long as the human is already skilled and experienced enough to do those complex tasks.  The problem comes for those who aren’t.

You see, learning from our experiences still forms the bulk of how we pick up new professional skills.  If machines are now doing the easy work that employees would traditionally hone their craft on, then it fundamentally changes the dynamics of how people learn at work.  Sure, we might train in a simulated environment, and whilst that has merits, it isn’t really any kind of substitute for a real-world scenario.

The risk of de-skilling

This is a potential risk for fields such as autonomous driving.  The early iterations of autonomous technology is likely to perform the ‘easier’ tasks for us, whether it’s driving on the freeway or parking the vehicle for us, with the human then requested to take control if a higher skill level is required.

I’ve written a number of times about the challenges involved in asking human drivers to regain control of the vehicle after they’ve been unfocused on the task at hand for a period of time, but there is also the real risk that without the ‘easier’ skills that we take for granted whilst driving, we will lack the practice required to develop the advanced skills that are being requested of us.

Such a situation is perhaps starkest in a driving scenario, but it is easy to see how it might also apply in many other situations whereby machines take on the easier tasks.  It underlines the importance of ensuring that humans are constantly learning and keeping their skills up to date.

A range of skills

Of course, the same is likely to apply in reverse as well, as machines that are trained only on relatively simple tasks, will be wholly unequipped to deal with any more advanced or unexpected circumstance.

Whether it’s humans or machines, if you segment and cordon off the kind of knowledge that each is required to have, then it presents a number of distinct issues around knowledge shortages.  This is perhaps starkest in organizations, such as Google and Microsoft, who have declared themselves to be AI first organizations.  By placing machines at the forefront, do they run the risk of de-skilling their human workforce?

It’s one of the many things that has to be considered as we begin to roll-out AI-technologies on a larger scale and begin to consider the way man and machine will interact with each other in the workplace.

Related

Facebooktwittergoogle_plusredditpinterestlinkedinmail