How Will Artificial Intelligence AI Impact Healthcare? Part 6

How Will Artificial Intelligence AI Impact Healthcare? Part 6

Some will say that the term Artificial Intelligence, or AI, is not new. In fact, AI goes back to 1950 (see graph below) to a research paper written[1] by none other than Alan M. Turing – yes, that Turing[2]!  Turing was key in winning the war in Europe by helping break the code of the first storable programable computer in history, the German Enigma cipher machine, using the new Automatic Computer Engine (ACE). There are some amazing books written on him.

I found a fascinating graph on AI published by Harvard University from origins to 2000 that puts things in context. You can read the full article at the link[3] below.

I believe the reason everyone is talking about it today is mostly because OpenAI released their chatbot, ChatGPT, for beta testing in November of 2022, and it went out of the laboratory into mainstream social media as freeware (for a while). Where were you when you first heard about ChatGPT? I can tell you I was in a Technology Room in the Social Media platform, “Clubhouse,” and I will never forget how long it took to have access to ChatGPT. Though I have had it for months and use it daily (not to edit these blogs- so errors are all mine), the fascination is only getting deeper as we now have to learn “prompting” to get the best results. According to many people I’ve spoken to, Prompt Engineering has risen to be one of the fastest growing career paths– just like HTML5 and Javascript programmers 10 years ago I suspect. Perhaps that is the transition path for these programmers – from HTML5/Java programming (now becoming obsolete because of AI) and into Prompt Engineering?

In looking over some old presentations on our company, we have been using the terms Machine Learning and Artificial Intelligence for nearly a decade in describing our business model.

We love big data in healthcare, and since I am not an actual provider of care, my contribution is/was how to gather, move, and process data to make it actionable to an actual provider. In the process of building care management companies, we helped build some amazing companies that are still giving direct medical care to patients.

We are not fans of having AI directly give medical care other than patient intake and a few Q&As to focus on the patients’ complaints. Apart from this generative-AI ad hoc query (rudimentary AI at best), unless it is triaging and providing diagnostic assistance to the attending physician, we believe the best use of AI is the overall processing and analysis of data. In most cases, we must assemble the data and teach the “machines” how to learn, then guide it when it goes off track (using that ability that only computers have) in an incredible scale to improve the convergence of information to both the patient and provider.

Our current lead project tells you, as noted in Part 1 and Part 2 of this blog series, that we intend to take live patient data along with historical data, mortality and morbidity tables, and back test to predict the range of possible (rank order) future outcomes with high probability. It is not only fascinating, but we also think we will change how consumers see their own life and provide them a new set of highly accurate tools to positively affect their own healthcare. This same process, applied in a de-identified manner, can provide amazing insight to ERISA companies in evaluating their employee/beneficiary risk profile and what can be done to improve it. The same logic can apply to private insurance, and, in fact, our business model being prototyped, can be shifted away from health and wellness to other behavior, such as behavioral health. Amazing things can be done with the algorithms and data analytics stacks we are designing.

I recently Tweeted on May 12[4] that “AI is going to change Healthcare like no other innovation (but maybe fax and pager) has changed healthcare.” I believe that more and more every single hour.

-Noel J. Guillama, Chairman