AI Can’t Be Your Doctor, At Least Right Now

Artificial Intelligence can be your Artist, your Composer, and even your Bird Song Translator. But at least for now, AI has no chance of being your doctor (check out the links below).

Try not to cheat by looking at the chart – which industry do you think is furthest behind in adopting AI? If you said Construction, gold star for you.

But look what’s next to the bottom – HEALTHCARE! The industry that’s supposed to grow 13% by 2031 and add 2MM jobs, according to the U.S. Bureau of Labor Statistics. Money is not the problem here.

So let’s diagnose this patient: We’re seeing lots of AI attempts in healthcare, perhaps most famously by IBM Watson, which spun off as Merative. AI funding in healthcare reached $10B according to POLITICO – 35% of all funding for digital health startups. But why no big wins?

Like many medical issues, the problem is systemic, focused around three key areas:

1. Even if it’s the right answer, we absolutely need to know HOW any AI model arrives at the answer. You don’t undergo medical procedures based completely on “just trust the algorithm.”

2. AI requires data to train – a LOT of data. Healthcare data is moving into the cloud, but every competing system handles it differently. What data that CAN be used creates incomplete training sets for algorithms, which creates bad outcomes, which perpetuates the spiral.

3. This spiral decimates TRUST. Doctors, administrators, regulatory approvers – all of these audiences need to get on board before AI will meet you at the exam table. If you’re a Radiologist you might be most vocally opposed – AI devices gaining FDA approval to potentially take your job beat all other approvals by 6X.

Healthcare AI might not be smart enough yet to solve its own problem. In fact, this patient might best be treated by a Marketer, not a Scientist. #ThinkLikeAMarketer

Need a marketer for your business? Come check us out!