As a healthcare provider involved with technology, I am frequently asked about artificial intelligence (AI). AI technology aims to automate (and replace) many of the mindless, repetitive tasks we all do.
Nvidia initially got their start in the gaming industry. Since then they’ve migrated to virtual reality, driverless cars, and are now expanding to multiple other areas, including healthcare.
I have to admit I felt a bit like a party crasher. The average attendee was younger by several decades and a serious programmer. However, the excitement was palpable. It felt a bit like the early days of Apple when the iPhone was introduced. Nvidia’s CEO Jenson Huang’s keynote had a line 5-across stretching the length of the convention center. . Although I didn’t understand all of the technical jargon, I got the gist. AI is going to have a huge impact on almost every industry. Not surprisingly, Jim Cramer from MadMoney featured Nvidia a couple days after I returned. Cramer lists 4 things propelling Nvidia’s booming business.
I’m not an expert in the space, but here are a few observations from a physician’s perspective:
Healthcare and AI
Demonstrations were not ready for primetime. Most talks presented solutions to non-problems, things I do perfectly well already. However, I could see many routine tasks I do over and over again going away. This would save time and possibly improve my job satisfaction. Over the next 3-5 years, I believe the biggest opportunity is in the clinical research space, but clinical care is coming.
Amazon and Tech Rentals
In addition to books, Amazon sells technologies allowing companies to more or less “rent” computers and technology. While in their booth, I met Aran Khanna, a pleasant and affable, 23-year-old with a recently minted Harvard Ph.D. in math. In college he took information from the press and essentially unmasked patient records based on date, age, sex, type of accident, etc. —HIPAA protects confidentially . . . sort of.
Khanna showed me a number of technologies Amazon is working on. When asked where the space will be in 3-5 years, he noted 2 main issues—further improvements in chip technology (a given) and expert review of data, aka annotated data sets. These data sets are used to train the AI computers. In fact, these annotated data sets (human knowledge) may become the currency for the AI industry. Doctors aren’t dead (yet). In part, they will be innovators (asking the right questions) and trainers for AI.
Dr. George Shih at Cornell created a new start up, MD.AI for annotated data sets. His company finished 6th in the prestigious Kaggle competition. In George’s opinion (and I agree), people will want personalized AI to help solve problems in their daily work activities.
CloudMedx was the most interesting application I saw at GTC. This AI start up out of Silicon Valley compares patients to a database of medical records (thus far they have analyzed 10 million patient encounters and are currently in the process of analyzing 20 million more). Doctors simply dictate the patient’s complaint via natural language processing. The application then creates a summary note based on the combined information. Once this information can be placed back into the medical record, healthcare will have the Holy Grail. No more clicking through Cerner or Epic. Doctors can go back to being doctors and not data entry technicians.
Google was a notable conference omission. With their expertise in “search” and their new effort to build microchips, Google is a worthy competitor to Nvidia.
AI is automating automation. There will be winners and losers. Many repetitive (mindless) tasks are going away. However, many people, including me, make some of their living on these tasks. I’m worried about my profession, and perhaps more importantly, my children’s future. Time to pay attention and figure out your role in this evolving economy.