Slaves or AI

Aristotle was famously asked if societies would always require slaves.  He said “yes” and then qualified his answer by saying that, perhaps mankind could eliminate the need for slaves if it could invent mechanical devices to do menial tasks.  But as he didn’t think such machines likely, he imagined mankind would continue to require slave labor.

Nominally, Western societies eliminated slavery soon after the industrial revolution, in the mid-Nineteenth Century, when machines vastly increased our ability to do work and produce products.  In fact, there are still lots of very unpleasant jobs that need to be done, and its common for critics of our societies to use the term “wage slavery” to refer to mankind’s need to regulate labor to accomplish menial tasks.

Today scientists and industrialists are exploring combining Artificial Intelligence (AI) software and mechanical automation to eliminate lots of jobs.  A part of mankind fears being put out of work and resists AI. Another part hopes to realize Aristotle’s dream and significantly reduce the risks and drudgery of millions of people by handing off all menial work to machines.

Some point out that, since the beginning of the Industrial Revolution, labor saving devices have always resulted, in the long run, in more jobs being created than lost. Optimists are convinced that, whether or not more jobs are created or lost, we can solve the distribution problems and seek to push forward.  I personally believe that change will occur, whether we say we want it or not, and hope it will be for the best.  It’s certainly an opportunity to eliminate backbreaking, repetitive labor by millions of people and to usher in a new age of leisure for millions more.

Those of us involved in business process change are close to the tip of the spear when it comes to technological change.  Change requires that organizations examine how work is currently done and then imagine how it can be done once new technology is implemented.  Proceeding to document and implement new work practices follow from new process designs.  Today we are being asked to help implement new AI-based practices.

As a practical matter, new AI technologies are just another round of computer automation. We are already familiar with what’s involved with implementing new ways of using computers. AI applications tend to be more comprehensive and complex, and require more sophisticated interfaces with human workers and managers, but otherwise AI is just more of the same.

For most business process people, our concerns will be focused on identifying the most cost-effective ways to implement AI applications.

Most of us are aware that “artificial intelligence” actually includes a wide variety of different technologies, from machinery to implement automation to software to solve problems and make decisions.  Similarly, some AI applications involve complex pattern identification, as visual systems do, and some apps involve gathering information from verbal inputs or textual sources, weighing the inputs and then making complex decisions.  Beyond the more focused uses of AI, combinations of technology offer more complex applications.  Combining vision systems, automated steering and breaking machinery, verbal input and output, GPS reading and planning apps, and real-time decision systems, we seem on the verge of creating self-driving cars.

Some AI applications involve rule-based applications that can explain their “thinking processes” (by displaying the sequence of rules used to reach their conclusions) but most new apps rely on neural networks that use complex “weighting” algorithms that defy human understanding. We train a computer using cases and lots of practice, and, after hundreds or thousands of trials, the network learns to respond to cues we may not appreciate, to perform tasks we value.  Some find this worrying, but for most of us this isn’t a problem; we routinely rely on the opinions of human experts whose reasoning we don’t understand, like our physicians or computer repair technicians, and we survive. 

The AI applications that are getting the most attention, at the moment, are the so-called “large language models” (LLMs) that involved neural networks with 10s to 100s of layers of nodes that are trained on problems until they seem “human-like” in their ability to talk about situations that are important to us.  There are currently a dozen or so of these LLMs commercially available, and each year they become larger and more numerous.

Computer software developers, like Microsoft, are working to embed LLM capabilities into a wide variety of the software we already use for day-to-day applications.  In most cases the embedded AI should make our routine applications easier to use and more flexible.  It’s easy to imagine that applications will increasingly conduct dialogs with us to determine what we want, and then to help us as we seek to personalize our applications.

Read the column by Howard Smith for a good example of how Southbeach, a popular application used by business process professionals to analyze complex problems has been enhanced by the use of a popular LLM.

One of the most popular applications of AI LLMs is to generate code.  In essence, the long held dream of automatic programming seems on the verge of at least limited fulfillment. One describes what one wants a program to do and the AI LLM generates the code in an appropriate computer programming language to perform the task.

In a similar way, new LLM apps are being explored that offer managerial support — as, for example, by interviewing and ranking potential new hires — or legal support — by gathering information and drafting contracts or helping prepare taxes.  And there are a growing number of tutoring applications based on LLMs that discuss topics with students and suggest alternative ways of approaching problems.  Kahn Academy, a popular, free, online source of  instructional courses has begun to provide an AI tutor with its instruction.

The key thing to note here is that these LLM apps are not replacing blue collar tasks, but replacing highly paid programmers, middle managers or lawyers.  In most cases the apps do not propose to replace but to support existing performers, but the effect, overall, is to reduce the total number of programers, lawyers or managers require by most companies.  Companies that report on trends to personnel departments have begun to suggest that AI apps could replace some 40% of the workforce in the course of the next 25-50 years.  Of course all such estimates are simply guesses based on vague assumptions.  The key fact, remains:  AI applications have the potential to make major changes in the way business processes are organized at most companies.

Those involved in business process management are ultimately involved in helping organizations become more productive.  We seek to redefine processes to reduce the human labor involved or to increase what a given human can do in a given period of time.  The best tool in our toolbox, for several decades now, has been computer automation.  And, today, the best computer automation tool, for most of us, is artificial intelligence.

Every business process professional needs to learn everything he or she can about AI.  It’s the future of business process management.  Perhaps we can actually realize Aristotle’s dream and eliminate human drudgery from the world.

Paul Harmon

Paul Harmon

Executive Editor and Founder, Business Process Trends In addition to his role as Executive Editor and Founder of Business Process Trends, Paul Harmon is Chief Consultant and Founder of BPTrends Associates, a professional services company providing educational and consulting services to managers interested in understanding and implementing business process change. Paul is a noted consultant, author and analyst concerned with applying new technologies to real-world business problems. He is the author of Business Process Change: A Manager’s Guide to Improving, Redesigning, and Automating Processes (2003). He has previously co-authored Developing E-business Systems and Architectures (2001), Understanding UML (1998), and Intelligent Software Systems Development (1993). Mr. Harmon has served as a senior consultant and head of Cutter Consortium’s Distributed Architecture practice. Between 1985 and 2000 Mr. Harmon wrote Cutter newsletters, including Expert Systems Strategies, CASE Strategies, and Component Development Strategies. Paul has worked on major process redesign projects with Bank of America, Wells Fargo, Security Pacific, Prudential, and Citibank, among others. He is a member of ISPI and a Certified Performance Technologist. Paul is a widely respected keynote speaker and has developed and delivered workshops and seminars on a wide variety of topics to conferences and major corporations through out the world. Paul lives in Las Vegas. Paul can be reached at
Paul Harmon

Latest posts by Paul Harmon (see all)