Artificial Intelligence, Workflow, Healthcare, and The Fourth Industrial Revolution

[The blog post is in preparation for my participation in the Friday, weekly@HealthStandards #HITsm tweetchat Healthcare & Artificial Intelligence: The Fourth Industrial Revolution, hosted by @Drexdeford.]

“A just machine to make big decisions
Programmed by fellows with compassion and vision
We’ll be clean when their work is done
We’ll be eternally free yes and eternally young”
Steely Dan – I.G.Y. (International Geophysical Year)

One of my master’s degrees is in Artificial Intelligence (MSIS, Intelligent Systems), so I’m naturally fascinated by AI’s potential, to use a well-worn word, to “transform” healthcare. However, as fascinating as AI is, it is only one of a large number of emerging technologies, from 3D printing to autonomous vehicles to blockchain, and on and on. The recent Foreign Affairs article and subsequent book, The Fourth Industrial Revolution, lists AI twice in a list of twenty-three technologies, though many of the other technologies, such as autonomous vehicles and robots, rely on AI and machine learning.


The Fourth Industrial Revolution is not about any individual technology, such as steam power, electrictrification, or computing (the first three industrial revolutions). The Fourth industrial Revolution is not about the Internet of Things, 3D printing, self-driving cars, artificial intelligence, or big data. It is about the interaction among all of these technologies. It is not about innovative products, but innovative systems. Wearing my systems engineering hat (my other MS degree) I will argue that the Fourth Industrial Revolution is therefore about processes and workflow.

Here are a collection of recent (all 2016!) articles about the Fourth Industrial Revolution, starting with the most healthcare related. You are welcome to drill down and read any of these pieces. But my point to include them here is so you can skim their titles, quotes, and notes to gain a quick, general impression.

Healthcare and The Fourth Industrial Revolution

In the following excerpt from the Table of Contents of the book, The Fourth Industrial Revolution, the words “profound,” “systemic,” and “megatrends” occur. Interactions among three layers — physical, digital, and biological — will push humanity into a new era. The difference between our current computer-based 3rd Gen civilization and this new 4th Gen civilization is as great as the difference between steam-powered and electricity-powered, or between electricity-powered and computer-powered civilizations. Think about the impact of moving between previous industrial eras, in terms of both material wealth and health, but also social and international disruption. This is why so many commentators, to the degree they buy into the concept, predict both opportunity and peril.


Again, The 4th Industrial Revolution is not about innovative technologies. It is about innovative systems of technologies. It is not about product, or even technological, innovation. It is about multiple, different, interlocking, and rapidly evolving technology sub-systems becoming part of an even larger, and way more complex super-system, a system of systems.


(I am also reminded of the third book in Neal Stephenson’s series, The System of the World, in which “modern” and “Western” ideas overtake feudal thought)

How do systems engineers manage system complexity? With models. Systems engineers gather data and optimize these models. These optimized models then drive system behavior. Then more data is used to optimize, and so on. In the old days, systems engineers sometimes gathered data with stop watches and clip boards. I did exactly this, when I built simulation models of patient flow. Today, the Internet of Things and Machine Learning are reducing time scales to collect and process data down to potentially mere seconds. And today, process-aware systems, orchestrating and choreographing system processes and workflow, also potentially cause things to happen in the real world, also potentially on the scale of seconds.

What are “process-aware” systems? These are information systems that explicitly represent, in database format, models of processes and workflows. The models are continually informed by data. The models are continually consulted when deciding what to do, say, or steer next. While process aware systems “introspect” they are “aware” not in a consciousness sense, but rather in the sense that they can reason with these models, in real-time, in response to their environment, to exhibit intelligent behaviors that would not otherwise be possible.

Today, the industry most adept at representing work, workflow, and process explicitly, in a database, and using this data to drive, monitor, and improve process and workflow is called the Business Process Management industry. However, process-aware functionality is now diffusing into other kinds of applications and industries, including healthcare information systems and healthcare.

What is the relevance of process-awareness to The Fourth Industrial Revolution? This 4th revolution is about system, not product, innovation. Systems, as any systems engineer will tell you, are made of processes and workflows. The 4th industrial revolution is about systems of systems. Each of the technologies on the list is a technological system, but the 4th industrial revolution is about putting systems together to create entire new economies out of interacting, co-evolving industries.

Why is BPM (Business Process Management) so relevant to creating and managing effective, efficient, flexible, and satisfying systems or systems? Because, figuratively, “WFM/BPM systems are often the ’spider in the web’ connecting different technologies,” and different technology systems.

Now I will turn to artificial intelligence, because it occupies a special role in The Fourth Industrial Revolution. In particular, machine learning will turn big data into big workflow. In other words, AI/ML will make big data big actionable, in realtime time and at scale.

In many ways, workflow technology, while not a direct descendent of early artificial intelligence research, nonetheless inherits important similar characteristics. Both distinguish between domain knowledge that is acted upon and various kinds of engines that act on, and are driven by, that updated domain knowledge. Workflow engines are like expert systems specializing in workflow. Just as expert systems have reasoning engines, workflow systems have workflow engines. Artificial intelligence and machine learning are critically about knowledge representation. Early AI used logic, current ML uses neural network connection strengths. Modern BPM represents knowledge about workflows, works, plans, goals, activities, resources, and so on.

Finally, many AI systems, especially in the areas of natural language processing and computational linguistics, communicate with human users. When I say “communicate” I don’t just mean data goes in and come out. I mean they communicate in a psychological and cognitive sense. Just as humans use language to achieve goals, so do some AI systems (some using text, but others using visual symbols and gestures). Communication between humans and workflow systems is rudimentary, but real. Workflow systems represent the same kinds of things human leverage during communication: goals, intentions, plans, workflows, tasks, and actions. These representations are, essentially, THE user interface in many workflow systems.

Artificial intelligence and machine learning will play key roles in coordinating intra- and inter-system communication and coordination. So will workflow technology. In fact, workflow models and their execution will increasingly be guided by artificial intelligence logic and machine learning networks. And artificial intelligence logic and machine learning networks will increasingly relie on time-stamped workflow event data to create and improve these logics and models.

To sum up, The Fourth Industrial Revolution is not about any one product, technology, or even system. It is about innovation in how multiple systems of technology come together. Artificial intelligence, machine learning, and process-aware technology such as business process management will, together, play a special role in gluing together these systems, so they can be fast, accurate, and flexible, at scale.

The Fourth Industrial Revolution

Here are my answes to the questions for the tweetchat, Healthcare and Artificial Intelligence: The Fourth Industrial Revolution and Healthcare. See you there!

T1: Amazon, Netflix and others use AI to predict purchases/spending habits. Can we do the same to better engage patients; drive better health? #HITsm

Yes! By the way, the “pipelines” collecting, analyzing, and acting on customer data are supremely process-aware, in effect workflow tech applied to data.

T2: Eliminating Waste; Flow; Transparency — “Lean” Concepts — How can AI support “Lean” Healthcare Delivery? #HITsm

Workflow! Workflow tech is a form of rudimentary AI applied to workflow, think “workflow expert system.” Also see my slides notes for Is Lean A Good First Pass For Healthcare Entities Looking to Explore BPM?.

T3: Watson — The Real Deal? Taking Too Long? Overhyped? What’s on your Watson “To-Do” List? #HITsm

All three! (BTW I went to school with the CMU grad student who went on to design Watson.) Another by the way, at it’s heart, Watson, as least the original one that won at Jeopardy, relied on workflow tech to manage interactions among language and reasoning components.

T4: Bots, bots, everywhere: Diagnosis, driving visits to clinic/hospital (or not) and more. Your “bot” thoughts? #HITsm

I’ll let @MrRIMP, my pet pocket ‘bot, speak for me.

T5: Which jobs will be most affected near-term/long-term— job losses as a result of AI? Or “plugging holes” in healthcare’s personnel shortages? #HITsm

From a previous tweetchat.

@wareFLO On Periscope!


This entry was posted in social-media. Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.

Post a Comment

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

You can add images to your comment by clicking here.