The Tech Translator Will be the New Accountant
The other day, I went to the dentist to get a regular cleaning. I was a new customer for this practice so I had to go through the whole rigamarole with the x-rays despite sending them files from my previous Dental office. After the x-rays, I was escorted to one of the patient rooms where I’m waiting to be seen by the dentist. The dentist comes in and moves this computer screen right in front of us. We’re both staring at x-rays of my mouth for a few minutes with two completely different perspectives. While she inspects my gums and teeth for health, I am studying their software.
As I quietly inspected the product’s design and features, the dentist finally speaks. She apologizes and says she is still learning this AI stuff since it is a bit new in this software. She appears to have some skepticism about the new AI features (i.e. bounded boxes that identified specific situations in x-rays of teeth). Meanwhile, I am chomping at the bit to chime in since this is the first time I have ever been excited to be at a dental office. I asked her if the AI algorithm is improving and being trained on specific data in their office. Is there a feedback loop? She thought it was a guarantee that the model automatically learns and improves. I recommended that she check into that and said this is definitely not guaranteed.
She continues with inspecting my mouth giving me some more time to collect my thoughts. Once she finished the inspection, I brought up the option to tweak the models to produce more false positives versus fewer false positives. If she is confident in her ability to correct the model, more false positives might be a good thing and you can override the models’ recommendations as needed. On the other hand, if you want a more accurate model then you might prefer a higher threshold for a recommendation. I said there a few ways you could approach this situation. She thanked me as she exited the room.
This situation is not unique to this one dentist. It is not unique to the dental industry. These technical conversations are popping up in all industries. Having a strong foundation of tech knowledge, or access to someone who does, is already critical and it will become increasingly important as “software (and AI) eats the world”. Every organization, large or small, should have access to a tech (and AI) translator to understand the software that runs their day-to-day operations. These translators will need to be embedded in every organization like the modern equivalent of an accountant who understands an organizations’ finances. Without a translator, your organization will undoubtedly make costly mistakes, lose to your competition who embraces software (and AI), and you will no longer need the actual accountant.
~ The Data Generalist
Data Science Career Advisor