“The ‘smarter’ machines get, the more and more jobs they’ll replace. Including mine. And yours.”
— Reader Comment
Martin Ford’s prize-winning Rise of the Robots (2015) underlies this three-part look at the prospects that drones, robots and cognitive technology will rapidly displace jobs and careers at all levels of the Big Audit model, and that the speed with which those learning organisms will surpass that of even high-judgment humans is dramatically under-appreciated.
The span of initial reader reactions included the expected denial and resistance -– “the obligatory weekly robot article” and “tired of this argument” –- part of the automation dialog since the dawn of the Industrial Age when artisanal weavers led by Ned Ludd burned factories and destroyed looms in England out of their anxiety over an uncertain future.
More substantively, this part starts with the comment (compressed and emphasis added) that “Automation has been here forever… Since we who are still here have adapted in the past, I assume the majority of us will continue to adapt in the future.”
That’s a reader stating the “induction fallacy” -– namely, that past conditions are not only expected to continue, but provide guidance as to the likelihood that conditions will not change -– both a common and instinctive human reaction, and in general quite misleading. (For those so inclined, see the insightful if eccentric Nassim Nicholas Taleb in The Black Swan (2007) and, better, Daniel Kahneman’s brilliant Thinking, Fast and Slow (2011).)
Not “This time it’s different” –- but “The future is already arriving”
To start, the MOTS skepticism needs to be dispelled. The fatal accident on May 7, 2016 involving a Tesla Model S in “autopilot” mode, where the car’s system was somehow blinded and unable to recognize a white truck turning across its path, will not slow down the arrival of driverless vehicles. Quite the contrary -– now unleashed, the genii of technology will not be re-captured. The discussion has already shifted from capability to liability and responsibility. It’s no longer “if?” — but “how?” and “how soon?”
Those still believing that transformative deployment of automation is still beyond the horizon are entitled. Only think how far the early stages have already advanced:
- At our daughter’s university, the automated system at the school library takes orders and delivers books in ten minutes, making anachronistic the old picture of students rummaging through the dusty stacks.
- Automobiles are parked in fully-automated, multi-story structures, stored and retrieved without the touch of human hands -– benefits include reduced space requirements and air quality improvements due to the elimination of drivers idling in search of spaces.
- The UK government’s agreement with Amazon to explore drone delivery moves far to solve the challenges of automating the “last few feet,” that contributed to the failure of bubble-era mis-adventures such as Webvan, where aggressive expansion and lack of appreciation of the grocery industry burned up hundreds of millions before its bankruptcy in 2001.
“Citius — Altius — Fortius”
— Motto of the Olympic Games
Further to the conclusion emerging from Ford’s book that the Big Audit model will be no less affected than any other sector, consider two examples. Both respond to one reader’s comment, that “inventory counts are a small fraction of total charged hours for new staff” -– which is true enough, except that sampling processes have throughout history been rife with undetected company fraud and substandard professional performance:
First: Decades ago as I was learning the profession’s business, the story was widely shared of a Big Eight firm’s comprehensive audit of the rolling stock of a Chicago-based railroad. Rail cars being revenue-generating only when in motion, a major part of the engagement involved a fleet of two-seat airplanes –- a pilot would locate the trains as they crossed the prairies and fly alongside, while a flight-suited staffer with a clipboard clamped to his leg leaned out the plane’s open window and checked off the serial numbers of the speeding freight cars.
Not so “Indiana Jones” today. Audit teams can post themselves trackside, and scan the RFID codes on the cars as they roll by. And in a couple of years, a robot will do the code reading, or a drone will fly next to the moving trains. Total human staff for the job, instead of the by-gone teams in the field, will consist of an algorithm designer and a machine operator, reporting to the audit manager for analysis and evaluation of this mechanically-gathered audit evidence.
Second: It was my challenge in an earlier world of traditional audit methods to help defend a Big Eight firm against shareholder class action litigation and a grand jury investigation, unearthing the fraudulent scheme of a publicly-held client with a large business in the fitting and finishing of large ships for both civil and military use.
The below-deck interiors of the vessels involved long corridors of identical crew quarters and other spaces. So it was efficient during construction that when a cabin was fully equipped, fitted and approved, its door would be locked and sealed off, to receive no further attention until final delivery.
The client, feloniously gaming the system by which it could receive advance contract payments based on the state of progress, would simply seal empty cabins –- and persuaded the overly-trusting auditors not to insist on breaking the seals for eye-ball inspection of the actual extent of work completed.
So -– in the environment of today’s evolved technology? The same protocol of sealing cabins once ostensibly completed could be followed. But a robot or a drone, “walking” the corridors with an RFID reader and a miniature camera slipped under each locked cabin door, could read the codes and “see” the individual bunks, desks, sinks and toilets that legitimate installation would have put in place.
In short, if the machines will be able to audit better, smarter and cheaper, is there relief from the compelling inference, that the accelerating learning capability of the algorithms will increasingly displace reliance on human judgment?
Ford thinks not. The robots think the same. Listen up –- and stay tuned.
Jim Peterson was a senior in-house lawyer with Arthur Andersen for 19 years, leaving in 2001 to pursue his own practice and to write about the accounting profession – in the International Herald Tribune and now on his blog, Re:Balance. His book, “Count Down: The Past, Present and Uncertain Future of the Big Four Accounting Firms”, was recently published by Emerald Books.