Over the past year, AI has definitely generated some fierce interest and stirred up some strong emotions, fueled by the media.
And with the almost eerie capabilities of Generative AI, showcased by ChatGPT in early 2023, it’s understandable that AI would cause some uneasiness among many.
But is AI really that “scary”? Or the unstoppable force to be reckoned with, as the flood of webinars and articles may lead you to believe?
Having been in technology for 40 years this year (and having had my electronic interaction with AI technology in 1995) let me cut through the froth and break it down to some pragmatic guidance.
It’s always helpful to start on the same page, so here’s where I’m coming from when I talk about “AI” in the context of this week’s column.
A brief overview of AI
AI, or Artificial Intelligence, is really a catch phrase for a whole arsenal of different technologies that are aiming to mimic human intelligence (and beyond). What has really caught the attention of the public is the power of Generative AI, enabled by leaps in computing power at declining prices.
But Generative AI is but one of many technologies. Here are a few terms that you may have heard of AI technologies that are real and in use:
Generative AI: Generative AI is technology that is capable of generating new content (or data, such as images, video, or music) based on trained data. It is currently one of the most advanced AI technologies, which is therefore catching the imagination of virtually everyone. Depending on the way you look at life, the reaction is either one of amazement or one of grave concern. The truth probably lies somewhere in between the two, but that’s a topic for another column.
Reactive Machines: Going back in time, this technology is capable of reacting to the current situation, but has no memory. The 1996 IBM Deep Blue that played Chess against Kasparov is an early example. The Jeopardy playing IBM Watson, by the way, was an early generative AI.
Limited Memory AI uses past experiences, but has no long-term memory. This type of AI is employed in self-driving cars.
Narrow AI (weak AI) is focused on specific tasks and Siri and Alexa are examples of that technology.
Expert Systems are custom-designed for a specific field of expertise to mimic the decision-making abilities of human experts, such as medical diagnosis, or financial advisory.
The next set of AI Technologies are the ones many scientists and policy makers are concerned about, as they present the next generation of AI technologies. At this point they all are hypothetical.
Theory of Mind AI (hypothetical) is a category of technologies, such as robots, that are capable of understanding and interpreting the mental state of other beings. Some attempts are made to look like this AI, but are mostly Narrow AI or Limited Memory AI solutions.
Strong AI or Artificial General Intelligence (AGI, hypothetical) are labels for AI that is self aware and could perform any intellectual task that a human can. That is the AI many are concerned about.
Artificial Superintelligence (ASI) or Super AI (hypothetical) surpasses strong AI in human intelligence and ability. This form is, however, still purely speculative in the absence of working Strong AI solutions.
AI – what’s in an Acronym
“What’s in an acronym? Would AI by any other acronym still be as scary?” — with apologies to Shakespeare
One of the most fun parts over the past years has been for me to collect different translations of that “AI” acronym (technically an initialism) into different meanings.
Beyond its original form of Artificial Intelligence I like these various variations:
Augmented or Assisted Intelligence. That sounds much less scary and actually quite helpful. Many of the AI solutions currently fall into this space.
Actionable Insights – that is also a new favorite of mine, especially when explaining AI’s value in medical diagnosis and treatment. For the time being, we will not purely rely on AI to make decisions for us, but it can give us pretty good insights that are actionable, once we reevaluate the suggestions using our human experience and expertise.
One very memorable play on words (or on letters, in this case) was Thomas Friedman’s foray into turning AI into IA, from his book “Thank You for Being Late: An Optimist’s Guide to Thriving in the Age of Accelerations”. Incidentally, he was the keynote speaker at an American Telemedicine Conference in 2017 or 2018 where I first learned about it.
Friedman suggests thinking of AI in the form of “Intelligent Assistants” as well as “Intelligent Assistance”, as well as Intelligent Algorithms.
All of this is to say that these technologies can provide us with a tremendous amounts of benefits, when used appropriately — as, by the way, has always been the case with any technology from TNT and the atomic bomb to Freon or supercomputing.
So if you are concerned about AI, just think “Actionable Insights” and “Intelligent Assistants”.
AI in Healthcare: What’s a Leader To Do?
There are numerous use cases for AI in healthcare which I’ll explore in a later article.
So what are the implications for healthcare, specifically for healthcare delivery organizations such as health systems, hospitals, health centers, surgical centers, and behavioral health agencies?
If you’re a leader in any of those organizations, what should you do about AI?
Well, here’s the good news, the Reality Check and the Fine Print.
The Good News
When it comes to AI, here’s the good news for leaders of healthcare delivery organizations:
1. You don’t have to “implement” AI. Meaning that as a healthcare organization, you do not have to worry about “buying” and “deploying” AI, so not much changes from how you are deploying and using technology.
2. You don’t have to develop your own AI solutions. Larger health systems, specialty care providers, and ancillary services may go down that road, but if you are running a standard clinic, health center, or health system, chances are slim that you need to worry about developing your own AI solution. If you’ve never commissioned the development of an App, you are unlikely to develop your own AI solution.
3. Your existing vendors will get better by embedding AI. The way that AI will be (and is already) entering your organization is through new features (and sometimes add-ons that need to be paid for separately) that use AI to optimize the value to the solutions’ users. EHRs are the first examples and many other technologies (including office technologies such as Zoom) are swiftly following suit.
4. There are new, powerful solutions coming, powered by AI. As laid out in the definitions above, the more powerful AI becomes, the more opportunities arise to offer something completely new. Right now we are in the evolution stage, but multiple rAIvolutions will be coming (and you and your organization got to be ready to embrace them).
5. AI solutions will be beneficial to your patients, clinicians, staff, and to your overall financial sustainability and clinical quality. While AI has its risks, the benefits of the computing power to mimic human intelligence reliably and consistently will, when applied correctly, lead to all the benefits listed.
6. AI will make your clinicians and staff smarter, more effective, more efficient, and more productive (and if done right, less burned out!). In healthcare, we have been dealing with a long-lasting crisis of burn out, staffing shortages, recruitment challenges, and retention decline. AI, when deployed and implemented correctly (not like your EHR rollout), can be part of the overall solution to address all of those challenges.
The Reality Check (the “bad” news)
And now for the “bad” news…
1. There’s an existing avalanche of innovative digital health solutions, followed by a Tsunami of AI-powered digital health solutions. Venture Capital firms have invested over $110 billion dollars into digital health startups over the past 10 years and investments in AI will be even higher. These solutions will be proven with demonstrated value. If our traditional health system cannot harness these technologies to reap the benefits, then other non-traditional companies (eventually) will succeed.
2. As an organization, you (i.e., executives, leaders, managers, clinicians, and staff) have to get great at deploying new digital health solutions in months, not years. This realization has now become imperative as this new wave of technology will not be stopped, and the integration of those technologies is not as trivial as technical deployment with some training. These technologies will change the way we work, the way we deliver care and you need to prepare your staff and clinicians for this change.
3. Telehealth is your Dress Rehearsal for what’s to come. If you have not even mastered telehealth yet, get moving. Telehealth, in comparison to what’s to come, is a very simple technology. But the principles of how to integrate these solutions (whether it’s RPM or chatbots) is similar – you need new workflows, technology, policies, training, support supported by change management and project management.
The Fine Print
To address the inevitable question by some of my readers — yes, despite all the superb potential benefits, AI introduces a whole new level of concerns around privacy, cybersecurity, and, new to the bunch, also ethical implications that policymakers and legal experts are still wrestling with.
My guidance? Do what you’ve always done: Work predominantly with trusted vendors and hold those trusted and especially any new vendors to the same standards regarding privacy and security as you’ve always done. Put the burden of proof onto them, to be comfortable that they have done their homework.
Because…
AI is not going away.
The Hype is real.
But so is the value and the potential of AI.
Want to discuss AI with me or lead a discussion with your leadership team? Then reach out to set up an exploratory call.
To receive articles like these in your Inbox every week, you can subscribe to Christian’s Telehealth Tuesday Newsletter.
Christian Milaster and his team optimize Telehealth Services for health systems and physician practices. Christian is the Founder and President of Ingenium Digital Health Advisors where he and his expert consortium partner with healthcare leaders to enable the delivery of extraordinary care.
Contact Christian by phone or text at 657-464-3648, via email, or video chat.
