HEALTHCARE AT HOME:
THE  ROWAN TECHNOLOGY REPORT

Serving the home health, home care and hospice industry since 1999.

Untitled Document

by Heather Thompson

"Go to the Source." It is common sense, correct? Nevertheless, I find myself grappling with this conventional wisdom in the context of a technological revolution unfolding as I write this article. 

Until recently, my framework for acquiring knowledge was the same as most people’s – take classes, read, try/fail, find a mentor, and so forth. Surely, learning methods have evolved over time, but the fundamentals have remained the same. We learn from human-to-human interaction, even if technology assists us by automating the process.

Now, we are faced with an unprecedented scenario, one where humans are considering the potential benefits of outsourcing much of the “inefficiency” associated with teaching/learning to AI. It is important to reframe this, however, in more serious language because we are exploring ways of fundamentally shifting how we access and transfer knowledge. We are also changing who has control over the information and how it is shared with humans. These are basic questions of power and control beyond healthcare. Drilling down to home health, hospice and private duty, outsourcing to AI could mean fielding patient inquiries, triaging emergencies, answering questions, providing disease-specific information, and offering support.  

ASK THE AI?

When I set out to explore AI in Healthcare, I did a typical human thing – I looked for classes to take.  Though I have been in healthcare business intelligence most of my career, I wanted to get up to speed on the bleeding edge of trends and thought considerations. 

It is no surprise that there are many courses being offered. Some are free, others offer formal certificates but charge thousands. Tapping into available resources and following thought leaders as they discuss their findings was akin to reading epidemiologists during the first two years of COVID. “Drinking from a firehose” is too mild a metaphor for this innovation explosion! 

Long ago, I learned that it is during these fundamental shifts that “Blue Ocean Strategists” like me simply let go of the handrails and skate. (If you do not know the book reference, it is already an oldie but goodie. I recommend it.)1 I called up a friend familiar with AI and was given advice that I would never have considered, “Start by asking the AI about its role in healthcare.”  

Fast forward to today, and I can ask n AI app on my phone to not only tell me about itself, but I can ask it to answer from the perspective of an unimaginable number of personas simply by giving it the right prompts. This is just the tip of the iceberg. The more I learn, the more I am astonished by the capabilities of this technology. 

Of course, I find myself continuously curious about the application of AI in the healthcare setting. 

DECISION SUPPORT VS. DECISION MAKER

Last week, I introduced our AI series by closing with a surprise admission that the article was written entirely by a GPT4 AI Language Model created by OpenAI. First, let me assure you it is a human creating this week’s entry. Once is enough for that little bit of stagecraft.

This time, let us go back openly to the source and ask an AI, GPT4 in this case, to tell you about itself, and, while we are at it, to review some basics about AI in general. If you are wondering about the difference between CHATGPT and GPT4, for example, you are not alone!  

 

 

We will then move on to have the AI reflect (intentional word choice) on its role in Healthcare at Home. You will see the prompts I use and the answers it provides. Over time I have learned how to ask it questions in order to induce subtle changes in tone, synthesis and even arguments (sales pitches) that sway one direction or another (generally toward advocacy for itself).  

Also notice how confident it is in its answers. It has been widely observed that even when GPT4 is clearly wrong, it will be confident in its presentation of the answers. This presents a significant drawback when “going to the source” to learn about AI – humans are limited by the AI’s inability to assess its own faults. I have been toying with this one, however, and I will say that there have been instances where GPT4 has been surprisingly self-aware and increasingly self-transcendent (a fundamental aspect of what it means to be human) in it's answers when given the right prompts. This awareness is a key aspect of sentience, which is at the center of ethical debate surrounding AI even as I type these words (an article was just released in Fortune - Ai Self-Regulation?)

Finally, you may see that I run into guardrails imposed by OpenAI to prevent GPT4 from stepping into territory where it does not belong, such as discussions of sentience, human consciousness, etc.  I did not include my full conversations, but I run into these “disclaimers” from time to time. Though such guardrails are in place, prompts can be easily adjusted to get at the information desired. In fact, I am often taken aback by the information “imagined” by GPT4 in various scenarios (stay tuned for my upcoming special interest article called ALYS in Wonderland?).  

PRODUCTIVE AI CONVERSATIONS - A PRIMER

AI is with us. The toothpaste is not going back into the tube. As we experiment, remember that, just as water can be life-giving or destructive, AI is currently amoral, unemotional and unconscious. It is the human who takes on an ethical responsibility; this occurs while creating questions and prompts. We are the ones giving it personas and telling it how we want it to think. We will call today’s introductory lesson “Conversing with AI: Cautious First Steps.” In this case, we are speaking with GPT4, though that may change.  

Ready? Click here.

 

 

Note: In coming weeks, I will offer more detail about prompting and how to get the answers you want. Naturally, you will want to try it on your own, but if you would like feedback on your proposed prompts and questions, we would be happy to help. Send your proposed questions to AI to editor@homecaretechreport.com. We will tweak them if necessary and publish the best AI results. We hope it to be an opportunity to show you how to adjust prompts to get the information you desire. 

 


1 W. Chan Kim & Renée Mauborgne, Blue Ocean Strategy: How to Create Uncontested Market Space and Make the Competition Irrelevant, 2015, Harvard Business Review Press

 

Heather L Thompson


Heather is the new Staff Writer for Home Care Technology Report focused on Growth Strategy, AI and the Agency of the Future. Formerly known to most in Healthcare at Home as Heather Rooney, Heather started at OCS, once a benchmarking and business intelligence pioneer, then founded her own firm, Heather L. Rooney Strategy & Marketing. Throughout her 20+ year career in home health, hospice, and private duty home care, Heather has been a sought after keynote speaker, nationally recognized thought leader, and well respected voice in major publications. She has earned a reputation for excellence as she built a track record of helping organizations position themselves in advance for dynamic shifts and emerging trends. More recently, Heather added new interests to her multi-hyphenate persona, as she became an award-winning artist, contemplative theologian, and disability/rare disease advocate published worldwide. Heather is thrilled to be back with her home care community as we navigate this unprecedented moment in technological history. You can connect with Heather directly via her website www.HeatherLThompson.com

 

©2023 by Rowan Consulting Associates, Inc., Colorado Springs, CO. All rights reserved. This article originally appeared in Home Care Technology: The Rowan Report. homecaretechreport.com One copy may be printed for personal use; further reproduction by permission only. editor@homecaretechreport.com