A.I.: The Basics

AI is all over healthcare, from assisting in diagnosis to evaluating new medicines, from allocating resources to triage. Sure, there’s enormous potential – there’s also big risks. At last fall’s National Work Comp conference AI was all over the exhibit floor….in recent surveys HSA has conducted we have seen a dramatic rise in AI-related comments.

What’s apparent from our conversations with industry execs is this: AI is…in the eye of the beholder.

While industry folks talk about AI’s potential, they readily acknowledge their understanding is superficial at best.

I asked Jay Stith, the brains behind HSA’s analytical work – he’s also worked extensively with AI applications in his work with HSA and on the national scale for disaster prediction and preparation – to give you, dear readers, a very brief overview of what AI is, how it “works” and where it might be useful.

At its core AI represents the culmination of efforts to infuse machines with human-like cognitive functions. The engine driving AI’s transformative power is machine learning – a discipline enabling algorithms to learn from data patterns. This not only facilitates automation but also empowers AI systems to continuously enhance their performance, making them dynamic and adaptable to evolving challenges.

This potential doesn’t come without cost. Once you decide to pursue AI, launching a competent AI system requires a lot of work:

• Determining what problem you want AI to address,
• acquiring the resources (money and infrastructure) ,
• earning management and staff buy-in,
• acquiring the talent to develop AI,
• assessing/cleaning up/revising the data used to “train” AI
• developing metrics to evaluate the AI’s output
• building the AI model/tool/program/etc. structure,
• adequately training the AI, and
• then…the dreaded implementation phase.

All while navigating the tricky ethical considerations associated with AI (privacy, ownership, algorithmic bias, hallucinations, and employee displacement) and the looming threat of increases/changes in regulations.

That said…safely navigating the path will lead to much improved productivity, clinical outcomes, and lower costs for all.

More specifically, stakeholders believe AI in worker’s comp can be very beneficial throughout the workflow –from the basics like increasing speed and accuracy across the board all the way to enhancing predictive analytic capabilities and most, if not everything, in-between.

What does this mean for you?

The potential is huge but be mindful of the arduous process to get there.