Artificial Intelligence Recruiting

In less than 10 years "intelligent" AI will be prevalent in every "recruitment system" resulting in an accidental culling of employees.

As recruiters continue to hire leavers the machine will learn. It will learn about the inefficiencies in humans hiring humans. It will learn which humans do a good job vs. those that don't. It will monitor all of our digital activities and determine who is engaged and who is not. Who will last and who will not. It will reallocate people into different roles based on predictive assessments of real time job performance. It will share workforce performance data with other machines in order to establish industry benchmarks swapping people across organisations. People will become simple pawns in a game of financial chess. Even the CEO will be assessed for their performance (thankfully).

The machine will look for roles that can be done by machines, roles than can be done by robots, the few roles that still need humans.

The problem is the machine can learn but it cannot think. A frequently discussed long-term goal of some AI researchers is to develop systems that can learn from experience to surpass human performance in most cognitive tasks. But will it be she who teaches the machine that will be putting her thinking into what the machine should do when it learns. We think the machine can only determine performance based on what it is told that good looks like but, and here's the rub, it will learn comparisons and make decisions based on those comparisons. Out-thinking humans.

Maybe this is a level of intelligence that humans cannot achieve as we cannot crunch that much data. Hence we use our gut. Hence we make so many errors. Hence the machine will be more intelligent than we are and can actually ignore certain elements of human intelligence. It will not only learn but think in an articificial, data driven, evidence based intelligent manner. Something we humans cannot do because we have hearts. Which is why we are human of course.

Our AI systems must do what we want them to do. Not what they "think" we want them to do.

Of course, when the machine is working it will go far beyond just hiring and will more accurately forecast workforce needs resulting in fewer jobs and less security. But fewer workers could mean lower costs and therefore lower product costs so things become more affordable and therefore the poverty level lowers and for some (many) they have (affordable) leisure time rather than simple redundancy.

Getting the machine to work will be more about writing the correct specifications which is often harder than writing the correct code. Designing simple AI rules for example, to determine what is high performance will likely require expertise from people that do not exist today in HR or Operations.

System security will in turn become ever more critical as a major competitor can use AI-based cyberattacks to out perform you more so by hacking your system than developing a better product. Change the algorithm to allocate too many people here, make to many widgets there, too few distribution people over there. Wreck the supply chain.

And where does this leave recruiters? They were the first the machine got rid of. It worked out within a few days that it could do a better job. Without any data! The rest of course is history.