The Atlantic recently published a great piece on the increasing role of data analytics in human resources. You should read the whole thing.
In summary: American corporations are developing algorithms that can uncannily predict a potential employee’s performance based on data they’ve generated. Think Moneyball, but applied to waiters, programmers, and even corporate management. Old metrics like previous experience and education are starting to give way to fine-grained data analysis that probes into extremely specific factors that can make an individual right for a job. And for employees who are already on the job, additional analytics can provide objective performance metrics in real time. If these practices continue to spread, it could change almost everything about the way companies and their employees interact.
The author of the piece expresses a hopeful view of the data-driven job market. Sure, there’s something unsettling about your future being largely determined by machines, but on the balance, the system will be untainted by human bias.
On its face, this seems like a fair assessment. Many of the things that currently affect the success of job applicants are beyond their control, and at times, utterly discriminatory. Things like age, height, looks, and race are undeniably factors in the world of gut-level hiring decisions. When oh-so-fallible human beings are given the reigns, they tend to pick people who are a lot like themselves, or fit into their mental image of what a certain kind of employee should be.
Maybe it’s time to trust the machine.
And yet, trusting the machine is hard. Even if dumb ol’ Homo Sapiens are biased and inefficient, there’s something that’s nearly impossible to accept about ceding so much control to a meticulously-researched piece of code. Once you read up on the data, it’s easy to see why computerized hiring could be great for a lot of people, but it almost doesn’t matter. It seems wrong.
This feeling is probably due to the fact that hiring based on impartial data analysis involves a shift in our entire perception of what it means to be human. Even if we don’t have a life that fulfills us, it’s always nice to think that in some sense, we’re free to change, and do whatever we want. The person across from the desk during your job interview may toss your application out because he doesn’t think you have enough hair, but he could also accept you for an equally irrational reason. That human element gives us the hope that there’s always a chance. Even if, in reality, that chance is virtually non-existent. Meanwhile, when an algorithm chews up your personal data and spits out a LOW POTENTIAL message, you don't even have a fool's hope to cling to.
The idea of having your behavior constantly monitored for performance clues is even harder to swallow. As with the previous case, there are plenty of good things advocates can point to. Analytics-enabled employees will always know where they stand, and can be given clear instruction on how to improve. Without hard data, performance reviews can be laden with human error. Bad employees can evade detection, while good employees can be maligned, based on faulty perceptions. The computer, on the other hand, only sees the facts. And once it shares those facts, you can improve! Everything works smoother than a greased wombat.
However, this raises a question: how much are we willing to give up for a fully optimized career? Constant data analysis could result in more productive, happier workers, and more profitable companies. But once again, it also involves completely altering our perception of human life. Once we view ourselves and others as sets of metrics, things start to change. Like pro athletes or Pokémon, we'll have to become obsessed with getting our stats up if we want to stay relevant. The fact that personal factors can have a huge impact on our job performance further complicates the picture. It's not a stretch to assume that data-driven optimization at work will require data-driven optimization at home.
(“Sorry kids... you’re really cutting into my weekly numbers.”)
At a certain point, it might be worth considering how much we should really value absolute economic efficiency. If the dreams of data scientists come true, and every company is full of employees who are able to generate the maximum amount of revenue-per-hour, what then? In some ways, people might be happier. In other ways, they might be less happy. And eventually, everyone will have to grapple with the meaning of life in ways that no piece of software can comprehend. Big changes in the future are almost inevitable. Whether those big changes are also big improvements in the long run remains to be seen.
No comments:
Post a Comment