On the Seedcamp podcast, Carlos Espinal explained why Healthcare was the most actively invested industry for AI startups last year.
Mundane tasks are prone to error
Think about how many things that smart people do — like pilots — and yet we make simple mistakes because we didn’t go through a checklist.
A lot of the tasks that are done in healthcare are mundane, which leads to boredom, which leads to error.
Not enough humans
But it’s not enough to think about this from a strictly process point-of-view. Taking a risk on automated predictions is more appealing when human systems become inadequate, especially when lives become at stake. Carlos mentions two Seedcamp companies:
Viz.ai goes through ultrasound and radiological images, and helps specialists in specific hospitals take faster action on image analysis without having to rely on generalists in radiology.
In this case, radiologists are becoming less available in the US, creating longer waiting times for X-Ray results and other gaps in diagnosis.
Creative ways to get the data
Predictive and categorisation algorithms require large datasets for training, and healthcare challenges often come with historical data, or people are willing to share data since it helps with their diagnosis:
Gyant has an app right now that is collecting information from people who believe that they’ve been affected by the Zika virus. By funneling the data through a triage protocol, they’re getting visibility and training their algorithms to better understand when somebody does or doesn’t have Zika. They’ll be able to roll that out to different types of point-type analyses, like STDs, so that medical practioners aren’t overwhelmed by some of these cases, especially in situations where there’s a mass breakout.
In some cases, AI relieves an endemic stress, like not enough radiologists, and in others, it helps react to epidemic spikes, like virus outbreaks. In both cases, AI isn’t as much a cost-saver as it is a way to augment a human system that’s overloaded.
Without having to do the same thing over and over again, AI can take over.
Strategic exercise: AI Readiness
While this isn’t a repeatable way to evaluate all industries, Carlos revealed 3 heuristics that are relevant to Healthcare— mundane, repeated tasks; available data; a paucity of expert humans.
Based on this, which industries do you think will be augmented by AI next?
This is a sample lessons from ai.source.institute — you can sign up for free today.