The perils of machine learning


The artificial intelligence of tomorrow will learn right from wrong based on our actions..our data. To train machines to be empathetic, morally sound and unbiased, are we modelling that behaviour? How much does the data fed to us affect our behaviour, and if so.. how are we different than machines?

There is no putting it softly,
I’m hardly perfect.

Going through the motions
with my preconceived notions.

My vision in my line of sight,
filtered to see black from white.

Know right from wrong, still I lie,
standing on my moral ground.

Is it fair that am an instance sought,
when data itself trains my thought?

To break a barrier, courage mustered,
Isn’t it regressing then to be clustered?

And when I disrupt to diversify,
A machine learns to classify.

Propagating back, decision untoward
Should my fears be fed forward?

A faster route for time to compete,
driven by thought in a passenger seat.

Survey monkey tricks for users elated,
I’m sorry I’m not a design con-templated.

When I am special, the world apprises,
my machine learns me and generalizes.

And all I love and share and seek,
Data mine is mine to keep.

The data points how bond compliers,
Aren’t stories written by outliers?

And I would converge, me a naysayer
If there’s a deeper learning of my hidden layer.

Learning from mistakes to make a new one,
Can I first be a model human?

Trained to stay in line and tested fit,
Are you righting code or wronging it?

Tags
design