Close×

In case you missed our regular hail column from our November/December issue, here it is

Hail.com’s Joel Knott has been armpit-deep in AI software development and says the technology won’t steal people’s jobs – just do the dull bits.

There has been a lot of fear mongering in the news recently about Artificial Intelligence making us mere humans redundant within our jobs. And worse – taking over the world. Most of us who worked in the automotive industry prior to 2000 could have never imagined the technological advancements within our respective fields that we have today, and it can be scary to think what is coming tomorrow.

When I picked up my first dent repair rod fashioned out of a 48” length of 1/2” spring steel and began pushing my introductory hail dent in the car park of our Hobbs, New Mexico workshop, I never imagined that in less than three decades I would be assisting Deep Learning Engineers in developing AI software that deletes the human from the quoting process of hail damaged vehicles.

So, it may be somewhat contradictory to use the words “fear mongering” and then “deletes the human” but please hear me out. For context, our company is primarily a hail repair company, but we have an arm of our business that produces software specifically designed for the hail industry. One part of that software is an AI program that produces a repair quotation after being scanned by our HAIL3D Scanner. It is this time I have spent in the trenches of AI that has both opened my eyes to the possibilities and has calmed my concerns about being replaced by an AI-powered robot.

In the 1980s, an Austrian Robotics and AI engineer named Hans Moravec conceived the self-named “Moravec’s Paradox”. Moravec’s Paradox is the counterintuitive observation that tasks which are easy for humans to perform are often extremely difficult for machines, while tasks that are hard for humans are easy for machines. Since that Mazda 323 bonnet in the Hobbs car park, I have looked at tens of thousands of hail damaged bonnets. But ask me to count the exact volume of dents broken apart into size categories, and I will take minutes to perform the task and likely give you a slightly different count each time. This is something our software takes seconds to do and is repeatedly accurate. Really no different to how a basic calculator can work out extraordinarily complex mathematical equations in a nanosecond, but that same calculator’s astonishing “brainpower” could never compute the nuanced steps involved in replacing and welding on a beaver panel, something that you could do without thinking about it.

After seeing under the hood of AI development, I believe it is something we can all embrace to improve the efficiency of our businesses, like the aforementioned quoting example. Quoting is a part of our industry that we all perform as a task for free. So, for my money, if I can make investment in something that not only produces a more accurate quotation, but also allows us to pivot our hail technicians off the floor providing a task for free into repairing cars that produce an income for the company – well, it is a no brainer. Do I think AI will ultimately replace panelbeaters, PDR technicians and mechanics? Not in my lifetime. 

comments powered by Disqus