Robo-HR: An Idiot's Guide to Hiring and Firing

I suppose with the incessant drive to introduce ever more smart technology to simplify our lives, it was only a matter of time companies began using machine learning to do the messy work of HR.  A slew of innovative tech solutions like Doxa, Gild and Textio are promising to do a better, more objective job of making hiring decisions by making sense of mountains of data (including data mined off the internet).  What's best is the promise of making sure that now only will the new hire be a better fit in terms of values, beliefs and work ethics, s/he will also be hired free from bias (e.g. gender, race, sexuality, socio-economic background, old girl/boy networks, etc).  That's where the debate lies.

Most comments I've come across don't seem too optimistic though.  Not because they don't support the eradication of bias, but because they don't trust computers to make such a complex judgment which most people feel comes down to one's gut feel.  They might well also be right that computer algorithms could encode the bias of the coder, and if that were so, machine learning can hardly be trusted to make an accurate and objective choice.

But I think that counterargument seems to miss the point.  Whether the proposed solution can accurately predict or make a good decision is one thing. But that's different from whether it'll be good to have such a solution in the first place, especially if eventually, we can make it work.  And to add to that, I also hope that companies can be a little more explicit about the way they design the algorithm so that at least there can be public accountability and reassurance that HR decisions aren't purely based on old boys networks, race, gender, ethnicity, religion, etc.  

Of course, people of certain social variables tend to have shared values and beliefs.  But I think the values and beliefs, and how they fit the organisation should be prioritised, not factors beyond their control.  And I also think governments and NGOs should publicly acknowledge and reward companies that do.  No one's going to do that on their own - yes, I know companies love to say how much they believe in diversity and how diversity makes economic sense but really, who knows?  At the end of the day, it's about dollars and cents.  Which sensible businessman would invest so much money into diversity for its own sake, without solid proof that it leads to greater profits?  If these tech solutions can prove that hiring decisions were free from traditional bias, and if governments/NGOs can back their use, may be this could work!