THE GRADUAL PROGRESSION OF AI INTO OUR PROFESSION IS NOTHING TO FEAR
“HEY, GOOGLE, will you write an article that I can submit for my column this month? … Google? … Fine, I will ask Alexa.”
“Yo, Alexa, if you are done scaring the nation with your insidious laugh, I need you to drop me around 600 words of amusing and prophetic prose that makes me sound both insightful and woke. Alexa … ?”
Ugh — I just realized I have neither a Google Home nor an Echo. I suppose I could do this the old-fashion way of using the last few neurons harnessed in my noggin.
The idea of AI expanding to simplify our lives is a novel idea. My original notion of Alexa was that of another distraction for my lazy kids’ generation. That was, until I heard my septuagenarian mother ask Alexa what the temperature was outside, and I realized that AI had, in fact, arrived.
GRADUAL PROGRESSION
I have witnessed a gradual progression toward more AI in our profession. Take, for example, the reliance we have on our pretesting equipment to aid us in the diagnosis of common and uncommon ocular anomalies. In fact, you can see this illustrated in ophthalmic company offerings.
The stalwart lensometer and phoropter are employed daily, and we don’t think twice about the manual labor used to derive the prescription. However, I can now walk in to my lane and use a digital refraction system, integrated with my autorefractor and EHR system, the screen lit with whatever Snellen chart I choose, and toggle between the lensometer, autorefractor and digital refractor I just garnered. Though not artificial, this system makes my refractions more “intelligent.”
No, really! There is an art to refraction that is intuitive to the clinician, and using a phoropter, which has a motorized prism compensator and a split cylinder lens, makes refracting astigmatism fun again. Moreover, I can apply my acumen to the refraction without even being in the lane. #Refractioninjammies
HAVE NO FEAR
AI is coming, peeps. In fact, as I have elaborated in other word salads, this is not to be feared. Rather, I ask you to embrace the notion of having devices that can provide more information faster, with a greater sense of reliability than previous devices and automatically integrate to existing technology.
Take, for example, the OCT. I will be the first to admit — see my confessions http://bit.ly/2hLuRQH — that seeing intra-retinal microvascular abnormalities is a career goal of mine. (I marvel at the heroic efforts of my colleagues who claim they see dot hems in a patient whose eye is only open the width of a pistachio shell.) Now, imagine you have an analyzer from the team at Google DeepMind (the name alone is cool enough for me) that taught a computer an algorithm to detect diabetic retinopathy and diabetic macular edema from retinal photographs. I won’t bore you with the trials unless you want to read them (http://bit.ly/2GMuGPC ), rather I will share that Google used 9,963 images from 4,997 patients, and the computer achieved 97.5% sensitivity and 93.4% specificity. Boom!
AI is being developed for glaucoma and AMD and I, for one, am psyched. OM