Artificial Intelligence.. the Weather and Art

Artificial Intelligence (AI) is epidemic. A host of mimicking software algorithms “learning“ from the volumes of recorded history will impact society - perhaps change it forever. We shall see...

There are many questions as to what those influences might be. Aldous Huxley had opinions that he documented almost a hundred years ago. Legislators in Europe are racing to get ahead of the technological curve.  Tech giants are sprinting to develop Artificial Intelligence applications and get them accepted into society before that happens thus making any legislation irrelevant and reactionary. Asking for forgiveness versus permission are very different things after the cat of AI is already out of the bag. 

2017 Was a Big Year for AI performing better than     
humans for the first time in several key categories. 

The materials used for learning include everything and anything that can be digitized. Many collaborative Artificial Intelligence algorithms team up within the huge and ever-growing black box of code to establish patterns from those enormous data sets. The software can appropriately reproduce these patterns upon request - more or less. 

Meteorological models are early examples of Artificial Intelligence. When I became a meteorologist in 1977 one of my managers told me to look for another line of work as my career path would be automated within the decade. Moore's Law from 1965 observed that the number of transistors on computer chips doubled approximately every two years. The computerization of society followed and it would appear that the impact of that digitization of knowledge has been exponential in growth ever since.

Meteorological simulations of the atmosphere leapt forward in precision and accuracy in both time and space during my career. A keen meteorologist in the early 1980s could easily improve on the barotropic model available as guidance in that era. By Y2K (the year 2000 when the world was worried about the coding of the date using just two numbers to identify the year), computer models of the atmosphere were the state of the meteorological art and they have only continued to improve in the last twenty years. 

The human hand analysis of weather maps was actively discouraged within the weather centre after Y2K. There was no time dedicated to that essential and very human effort of drawing lines and deciphering coded data into meaningful patterns. The hand analysis synthesized the wealth of coded data and human experiences into a personal understanding of the current weather situation. The important weather concerns of the day were discovered and the team got to work creating products to alert the clients affected. Producing an accurate forecast requires understanding the important meteorological forces at play. 

An AI limerick created up on the request
of my fellow meteorologist and
 friend, Jim Murtha

By the time I retired in 2011, computer simulations of the atmosphere were good enough that meteorologists could maybe not even bother conducting an independent analysis and diagnosis in order to produce a prognosis. Available time on shift was increasingly restricted and very limited for that effort anyway. Proper measurement of that forecast performance is still elusive but that is another story based on the scales of time and space as well as forecast lead-time. That Performance Measurement research of mine never got implemented... 

It took 30 years for the forecast made by my first manager to verify but his recommendation to find other work was still apparently correct. I became an artist...

My concern is that brain "muscles" only develop with some heavy lifting. If all you are doing is repackaging AI, then one can expect to become feeble-minded. Apparently, Aldous Huxley felt the same way. 

Vincent van Gogh art "created" by
https://creator.nightcafe.studio/

The nested algorithms of AI can reproduce the genius of Einstein, Van Gogh and Tom Thomson with a keystroke. There are reasons to be alarmed and very skeptical and perhaps disbelieving. What indeed is real and what is the product of nested algorithms feeding on all of human experience? That tipping point happened a couple of decades ago in meteorology.

There are options though...

In art, the solution is to take the path not only less travelled but never trod upon before. Paint on and dam the algorithms. Maybe not even post your art because once you do, it can be assimilated by AI.

In meteorology, focus on looking at the real world. The techniques I describe in my Tom Thomson blog can be applied to any skyscape to better understand the real clouds and the actual weather. It is also fun to be outside surrounded by nature. Become a "weather walker"... 

Artificial Intelligent "Painting by Tom Thomson" is 
fortunately not even close (yet) but it tried...

It is time to re-read Brave New World by Aldous Huxley published in 1932. Aldous was right. As Huxley remarked in Brave New World Revisited published in 1958, the civil libertarians and rationalists who are ever on the alert to oppose tyranny "failed to take into account man's almost infinite appetite for distractions."

First Edition Cover from 1932

Seeing is believing is no longer possible with Artificial Intelligence. Beware of geeks bearing gifts... Boldly go forging your own unique path through life while enjoying that ride - it might be the only thing you can actually believe and appreciate. 

Warmest regards and keep your paddle in the water,

Phil the Forecaster Chadwick

PS: AI based on history often fails in situations that have not yet been observed. The now historic Progressive Derecho of Saturday, May 21st, 2022 was a bust forecast that impacted roughly 15.6 million people, representing about 41 percent of Canada's population. Strong winds downed over 1,900 hydro poles, five metal transmission towers, and countless trees along the path. There were 5 fatalities. The EC Warning Program had challenges.