image1 image2 image3


The rise of algorithms

Everyday, more algorithms creep into our day to day interactions.

The list of people you see first when you open Messenger, algorithm. The Instagram stories that you see first, algorithm. The list of products you see when you open Amazon, algorithm. The list of shows you see when you open Netflix, algorithm. The approval of your loan application, algorithm. The screening of your job application, algorithm. The buy/sell decision on your managed stock portfolio, algorithm.

And this is only the start.

Algorithms are coming to a lot more things from driving your cars to determining when and where and what seeds to plant next (farming and agriculture).

The rise of the algorithms has brought up concerns about the biases in these algorithms. Be it an algorithm that marks way more black people as unattractive than it does white people, or one that screens out more women than men from job applications, there is evidence that algorithms are and can be biased.

There are also questions around their accuracy. Like in the case of a self driving car that caused an accident, or the one that marked an iconic war photograph as child pornography.

The people that raise these concerns are making a case for how biased and inaccurate algorithms could be and why we shouldn't embrace them so wholeheartedly.

But the point these critics are missing out is that of how these systems would perform in the absence of these algorithms.

Rather than evaluate the algorithms against a maximum possible accuracy and lack of bias, they need to be evaluated against the status quo.

When we do that, we see that algorithms, though biased and not always accurate, are still less biased and more accurate than humans making the same decisions.

As we hand over more of our daily decisions to algorithms, we need not fret over them being biased and inaccurate. After all, if we were to make those decisions ourselves, we would be more biased and less accurate at it.

Just like we can begin to learn to overcome biases as and when we come to know about their existence, algorithms constantly get better as well.

I already hand over a lot of my decisions to algorithms - what to cook, what to wear, what to read, what to watch, etc. And will continue to do so in the foreseeable future.

Share this: