As industrial engineering evolves into “digital engineering,” will the automated factory work for the workers?
It is not a new phenomenon. Automation within manufacturing has long caused concern as to how it replaces assembly work that had been the domain of humans. Today, this feels like a familiar story, but actually we are about to embark on a whole new chapter. It is not just about the next generation of automation and robots, but about the AI (artificial intelligence) capabilities that come along with them. We are already seeing the genesis of AI in assembly factories, both in machines and factory management software. There will be effects, changes and consequences. However, rather than focusing on and accepting that the number of people working in manufacturing may diminish, let’s find a smart way to make the automated factory work for the people.
What we once called the third industrial revolution, which now we should refer to as Industry 3.0, has brought all kinds of automation to our assembly processes. Though it started many years ago, the momentum of automation has not let up. We see more advanced automated processes today with so-called robots, or “cobots,” closer than ever to resembling android expectations as defined in science fiction from authors such as Isaac Asimov. Throughout the years, as technology has progressed, the scope of work that could be automated has also progressed. If we look at what automation does today, a great deal of it simply could not be done by humans. Who would want to place thousands of SMT materials on a PCB by hand, with some materials the size of dust particles?
Automation is a key enabler for increasing product functionality and “sexiness”; no one wants a “brick” in their pocket to send a Tweet. Inevitably though, the challenge has been to find alternative employment for the sometimes large groups of people affected by the introduction of automation.
Extrapolating the progress of technology going forward a number of years, we see the digitalization of manufacturing, incorporating AI, bringing a similar challenge, but in different areas. If we do nothing, it is likely to feel much worse for those affected, as not only is the scope of those affected wider, but the pace of change is going to be faster, certainly making it more of a revolution, rather than evolution. Social impacts of technology shifts can be absorbed by society when they come at a pace where retraining or other societal shifts can occur to match in time. However, at the rate things are about to start occurring, the social reaction may not be able to occur in time to avoid serious ramifications for those most vulnerable.
The Human Value
Human operators possess three key attributes with which machines and AI software are in competition. People are flexible (i.e., dexterous), can easily communicate, and are thoughtful, all in highly evolved and sophisticated ways. It is an easy mistake to believe production operators willing to perform seemingly endless repetitive operations in manufacturing lack any degree of “cleverness.” Any professional production operator will demonstrate an extraordinary level of wisdom and capability relating to their work. Even an average operator is able to notice the finest detail of variation, with the ability to routinely adapt to changes in the environment, and they are able to literally shout about serious issues if necessary. Even simple tasks can be challenging in automation. For example, an operator need only be told a screw must be inserted, whereas today’s robots need to be clearly instructed how the screw must be picked up, traversed to the site, oriented, etc. Human skills are the envy of the digital robot/AI world right now.
The Robot Insurgence
Robots, however, are rapidly making progress. Robot dexterity is moving ahead due to development of new types of sensors and AI software that can comprehend the immediate environment in real-time. Now being able to touch and feel things that they come into contact with, as well as sense the pressure they are exerting, even being able to resolve visual data, makes robots more adaptable, safe, and more like humans. This technology is in its infancy today, as sensor technology and application develops, though the real limitation on growth is more attributable to AI digitalization. Software to make sense of all the necessary data in real-time, so that actions can be taken or avoided, is rather expensive to develop, so is limited currently to specialist applications only. “AI,” as we refer to it for manufacturing, is not like the visions of “Skynet” in the movies, a sentient machine we spawn one day and then argue for generations as to whether it has a soul, etc. AI now, and for some time, is more purpose-built, adaptive, heuristic algorithms with somewhat narrow purpose, but nonetheless very powerful and cognitive.
In terms of communication, we would imagine a huge gap between humans and machines in assembly manufacturing. Not so. This skill actually seems as limited in humans on the shop-floor as it is for machines. Humans can physically talk or even shout, but need technology to make their voices heard beyond their immediate area. There are no walkie-talkies, mobile phones or email, however, in most manufacturing operations today. Communication in manufacturing has been one of the key “pains” in the industry for quite some time, relying on direct human to human (operator to supervisor to manager to engineer, etc.) interaction through meetings, by which time information is old, the detail forgotten and is therefore mostly useless. On the machine side, we see no coherent mechanism yet for communication that includes the definition of standard for a common language and protocol across different types of machines, and machines from different vendor platforms.
In both the automated and manual assembly domains this will soon be put right in the form of the new IoT communication standard for the industry, the consensus-based IPC Connected Factory eXchange. The values and advantages that CFX brings for assembly manufacturing as a whole are significant. The depth of support for the many different types of automated processes enables large amounts of live data to be qualified and processed through CFX. Data collection from manual processes, however, is just as important. Utilization of CFX also encourages the digitalization support of manual processes, enhancing precision, mistake-proofing, traceability and control.
Finally, there is thoughtfulness. The availability of data from, for example, CFX is critical fuel for AIs and humans alike to have visibility and make decisions. Humans need the data to be converted into dashboards, charts, alerts and reports, and then, after some interpretation and discussion, perhaps decision and action. AIs being developed will perform that process in an instant, making them tools of surgical precision for the continuous tuning and optimization of the whole of the assembly manufacturing operation. Human reaction time will be seen as sedentary at best by comparison. The assumption in all of this, of course, is AIs will be as clever as experienced human thought processes. This is a tall order by today’s AI standards, and likely to be true for some time, though perhaps not forever. Until then, we will see elements of AI being introduced in stages as aids to human decision-making, following principles established by Industry 4.0. Examples such as Lean materials, finite production routing and planning can already be enhanced “intelligently” by digitalization, as can individual machine, group and line optimization.
Repercussions and Benefits
The effect of Industry 4.0 bringing AI to replace workers in our factories is no longer confined to the operators on the lines, but applies also to supervisors, managers and engineers. It would be easy to see how the concept of Industry 4.0 is being perceived as a threat by many working within the industry. The idea of doing yet another round, and indeed another level, of moving jobs out of manufacturing is not a prospect to which those involved will ever look forward. The question is whether this in fact will be necessary, or whether there are ways in which it could be avoided, or at least reduced.
It often surprises people to hear the original intent of Industry 4.0 was to help local small- and medium-sized enterprise companies. The focus is intended as a business principle driven by technology. The human resource in manufacturing is a critical asset, especially the accumulated know-how and experience. This resource is going to be needed, and be extremely busy as we transfer manufacturing truly into the digital age. This needs to be included as part of the business case, in a positive way.
A critical principle behind Industry 4.0 is the focus on the flexibility of factories to be able to effectively “make to order” with the same levels of productivity as if producing high volumes. The live management of the factory through digitalization and AI is the key enabler. Old-school engineering would call it “changeover management,” but AI takes it to another level on a continuous factory-wide basis, a truly holistic version. This capability to be flexible means factories can produce a wider range of products, without performance (i.e., business) penalty. Taking the idea forward works best with an EMS model. Rather than having factories dedicated to a single manufacturer and product line, EMS factories can be interleaving work-orders of completely different kinds of products, even competitive ones. Today, however, the typical EMS setup is line-by-line, one for each customer and for each product. With Industry 4.0, there needs to be more flexibility, less dedication. The integration of work-order execution, with on-demand choices of process configurations, yielding efficiency improvement opportunities as the spread of demand volume changes between products, needs to be managed throughout the factory by AI. Apply then this model to multiple onshore factories, sited local to the market and customer, rather than the existing huge-scale factories overseas. The all-inclusive cost to transport a product from a remote overseas manufacturing location to the customer is roughly between two and five times the cost of actual manufacturing itself (ignoring material cost). The saving on distribution costs, including the investment in the finished goods along the distribution chain, is compelling. It contributes a large part of the justification for investment in local manufacturing. Here then is the opportunity for existing local manufacturing operations to radically expand their operations, if armed with the tools of Industry 4.0. There will be many examples of local Industry 4.0 factories becoming more competitive than what will become “dinosaur” factories in remote locations. The savings made on distribution more than compensates for the labor cost differential, given that significantly less labor effort is required per unit produced. Many existing jobs being carried out by humans will be replaced by automation, but with the expansion of the business, many of these people can be retrained as the factory grows.
Dealing with automation is one thing. AI is another, as it affects not just operators but engineers and managers. We have already seen a gradual decline in industrial engineering expertise in our factories. What we have to remember is a significant portion of the core “nuts and bolts” knowledge of manufacturing technology that once required a deep understanding of industrial engineering has also changed as automation and software have been introduced. Key technologies and principles of industrial engineering are now executed within automated machines and, of course, within good manufacturing management software. The requirements for the human engineering/management team to run an Industry 4.0 factory of the future aided by AIs are quite different from running the old-school factory. Resources in manufacturing dedicated to getting the most out of automation and software must not be overlooked in order to save indirect expense. Automation and digitalization with AIs need to become recognized as “direct” contributors to the business. The old measurement of direct versus indirect labor will become obsolete very quickly. Getting the best possible result from automation, software and AI systems requires industrial engineering to evolve into “digital engineering.” Many cases of “digital stagnation” already exist in some manufacturing operations, with lost opportunity as a consequence. It is no longer necessary to understand nuts and bolts, bits and bytes of modern manufacturing technology. People who know about production flows and requirements need skills in the use and application of software. This is another opportunity to retrain and reassign those today with production knowledge and prepare them for the future. Having an understanding of operational requirements, and automation with system capabilities and functions, will be critical to making the Industry 4.0 and AI business plan work.
One would hope that converting jobs for those people currently “working to live” into higher-level more creative jobs may drive the change forward positively. During this transition, help will be needed from governments and sponsorship to make the automation/AI changeover work in as positive a way as possible.
Expansion of local manufacturing is not going to happen exactly at the same pace as increasing application of automation. No single solution will solve the whole problem, and many people may choose an alternative path, but the local manufacturing business case is becoming increasingly real and viable, if we choose to support it.