The obsolete geoscientist

The obsolete geoscientist

Andrew Pethick, Lecturer, Department of Exploration Geophysics, Western Australian School of Mines,

I had my first summer job as a junior geophysicist when I was 18. It consisted of processing potential fields, electromagnetic, resistivity and IP data for mineral explorers. I soon found myself repeating the same or similar tasks over and over. I was going mad! Rather than putting up with the grind, I coded up tools in C to speed up my work. Instead of spending a full 8 hours on a task, the job was completed hours ahead of schedule.

Now I had a dilemma, do I (i) tell my boss what I have done and tell him to send more work my way, (ii) Use the time saved to develop more time-saving tools or (iii) take a long lunch and tell nobody. Whichever choice, the job was done using the process outlined by the company and in the mandated amount of time. All victimless! I opted for a combination of options (ii) and (iii) only to go for (i) a few weeks later.

Fast forward ten years, a large part of my job is coding solutions for automating new geophysical processes. The way I see it, all software development is some form of automation or task simplification. From a macro point of view, software development attempts to shave a few hours off at the end of the day. Instead of everybody enjoying a long lunch, we just removed hours from someone’s day, or worse yet somebody’s job. Are we playing a prolonged version of musical chairs, with every played song a geoscientist loses their seat?

A key development area in both universities and companies globally is automation and machine-learning. We are (not just in the geosciences) working towards replacing human labor. We are pushing our work to be either very low level, to be carried out by the best and brightest developers or at the very high-level, to be carried out by the best and brightest leaders. My suspicions were confirmed by a high-level tech employee of a company leading the field in mining automation. When asked, ‘does the future mining company look like a select few industry experts surrounding a real-time ‘war-desk’ controlling fleets of mechanized drones?’, they answered, ‘what makes you think we don’t have this now?’.

There is nothing worse than seeing a man lose his purpose. Without geophysics work is there a need for geophysicists? Are we geoscience software developers working towards our own and our peer’s obsolescence; heading towards the Twilight Zone (1961, Obsolete Man) believing “no man is obsolete”?

Perhaps I am being too pessimistic. Yes, new technology opens up new avenues and possibilities. Yes, the job I am doing didn’t exist 20 years ago and yes, the world will keep turning. I am excited by what the industry will look like in another 25 years. Will we have 99.999% accurate automated first break and velocity picking AI? Will we have a robust unified inversion framework; data goes in, geology comes out? My only hope is that can contribute to any future industry revolutions in some small way. These software achievements are probably not unattainable pipe-dreams. Rather, the hardest challenge to overcome is human. Can an utopian AI-run industry be realized without discarding the human element? Possibly, but only if the tin man finds his heart.