The period 1714 - 1830 was one defined by balance and order, at least from an architectural stand point. One of my favourites. The four Georges who ruled during that time gave their name to large gridded windows, high ceilings, regulation terraced houses and the occasional dramatic series of columns. Georgian architecture is my field of wheat to run through with Theresa like joy.
For all my enthusiasm in this, our year of 2019, we live in more than just four walls. We live in our machines too. Having worked in start-ups for a number of years and then in tech education for a few more, I’ve built up quite a good understanding of what a modern day architect looks like.
I distinctly remember sitting on a sofa in the reception area of my last start-up (waiting for my final stage interview) and having a long haired, slightly dishevelled man with large headphones sit opposite me. He removed one of them and gestured as he said “Don’t mind me, I’m just trying to get in the zone.” I then watched him for about ten minutes tapping away, frowning, sighing and then tapping some more. A quintessential coder. A white dude in a band t-shirt.
Nothing wrong with band t-shirts of course (I have several) but if bricks and sweat built the world we live in now, then code and coffee are building the future. Are these buildings being built to house everyone? It would seem not.
This year US Congress saw the introduction of the 'Algorithmic Accountability Act of 2019' put forward to try and stem the deep bias that lives within the structure of most AI. To build artificial intelligence you need data. The machine learns from this data using the parameters of code it's built with and develops new information and understanding.
The issue is, the majority of data being used is an average and often this average excludes minority groups. Guys in band t-shirts build things for guys in band t-shirts. Facial recognition software for example - the kind in your new iPhone or Snapchat filter - has difficulty picking up darker skin tones and as a result differentiating between facial features.
Joy Buolamwini, founder of Algorithmic Justice League, told the story - during an oversight committee meeting - of a man who had been arrested multiple times by a retail store who used facial recognition software to identify wanted thieves. The software couldn't differentiate between him and another black man. Fortunately, charges were not brought but it begs the question; what if it had been a murder case?