Artificial Intelligence has gone from sci-fi to common talk in a very short time. Early adopters are leveraging it, while careful adopters still wonder where things could go wrong.
In workers’ compensation, AI can feel like the shiny new buzzword everyone’s talking about, without a clear understanding of what it actually does or what risks it brings. In a world where we need to instill trust, AI can be fraught with skepticism.
Why the Hype?
It’s new. It’s fresh. It’s intriguing. It is already unlocking new potential with promises of so much more to discover. Early success in our industry includes:
-
Faster review. AI can index documents faster, allowing reviewers to act more swiftly and removing delays in treatment.
-
Smarter claim suggestions. Summarization of medical records surfaces the most relevant information. It can spot patterns in claims and recommendations.
-
Consistent Guideline application. The same rules and guidelines can be consistently applied.
-
Error catching. Imagine a digital safety net that flags mistakes before anything goes out the door.
What are possible Hiccups?
The pace of evolution is overtaking the ability to test and validate its use. By the time you are comfortable with it, it has already changed. Trust has not yet been earned over time. For most people, it is easier to imagine the downside, particularly in a risk-adverse industry that thrives on anticipating what could go wrong.
- Security worries. Where is AI accessing data and how is it using PHI?
- Robot decisions. How was it trained? No one wants a computer to decide treatment plans. It is never ok to replace compassion and care.
- Unpredictable results. If AI isn’t set up or explained properly, it can create more confusion.
Responsible Adoption
We are responsible for optimizing processes and reducing costs, regardless of whether we are talking about streamlined automation or using AI. With AI, there must be controls to assure we leverage the good while minimizing unintended consequences.
First step: Don’t rely on generic open-source tools. At EK Health, in partnership with our sister company, we build models based on secure, proprietary data—not web content, keeping everything compliant and protected.
Second step: Define appropriate use. AI can speed up paperwork without creating a problem. But should it be used in clinical decision making? Do you want to turn your healthcare decisions over to a computer? We think through these considerations any time we deploy an AI solution.
Third step: Be transparent. Be clear about the scope of your use of AI. The best way to avoid unnecessary bias is to be clear on the limits of its use. There is a significant difference in risk between sourcing information for a UR physician to review verses making a decision typically trusted to a credentialed professional.
Care requires a human touch. AI isn’t magic, and it’s not the villain, either. Used wisely, it’s a valuable tool towards efficiency and expediting care. At EK Health, we believe you can have the best of both worlds, the efficiency gains of AI coming alongside the expertise, judgment, and compassion of real people.