If full automation were technically possible, would it be legally achievable, ethically justifiable or economically sensible? The race between men and machine is on. Dr Gernod Dittel explains
In the cleanroom sector, which is already highly automated, digitalisation is considered a potential job killer. The fear of full automation cutting employment must be weighed up with what is best for the cleanroom. Though automation may seem like the holy grail of cost-efficient production, many cleanroom experts believe man is superior to robots in many aspects.
For decades, humans have been developing cleanroom technology with machines slowly taking over more and more process steps. They carry out these steps faster, more accurately and, above all, cleaner than the human workforce. So should full automation be employed if possible?
The human being is the cleanroom’s greatest source of dirt and germs, an unpredictable and incalculable risk for the entire production line. The worker contaminates the end products and this can cause an expensive halt to production that requires cleaning and tests before restarting. This halt can last anywhere from hours to days, or far more. As a result, many people think that advanced process technology makes people in cleanrooms superfluous, and to an extent, this could be true.
For example, a few years ago in a circuit board factory in Shenzhen, China, 3,000 employees worked at cleanroom workbenches. After installation of a modern machine park, only three were left, with only one per shift working as a supervisor of the automated enclosed cleanroom production. However, despite the high level of automation of the process, there was still the need for these three people.
In some companies, the removal of the human aspect seems to be more within reach than in others. The management consulting firm McKinsey expects that in the US, 73% of jobs in the food, hospitality and hotel industries and 59% of manufacturing jobs could be highly automated. Another consulting firm, the Boston Consulting Group, sees 300,000 jobs disappearing in German industry by 2025 due to advanced robotics, which includes interactive industrial robots.
This automation revolution has been dubbed “Cleanroom technology 5.0”, and people are either pessimistic about its implications for the labour market, or optimistic about the promise of better quality for their products.
“Digitalisation is considered either a sword of Damocles, you’re afraid of it or the other way around, they say it might solve all our problems,” says Josef Ortner, Managing Director of Ortner Group in Austria.
The human being is the cleanroom’s greatest source of dirt and germs, an unpredictable and incalculable risk
The negative side is clear: the loss of many jobs. However, the optimistic believes there is a silver lining to the change. Two bestselling authors from MIT, Erik Brynjolfsson and Andrew McAfee, are saying that although it is an indisputable fact that the list of extinct professions may be long, the list of new ones emerging is even longer. This has been seen in the precedent set by the industrial revolution.
Jobs that were lost in the past were replaced by new, better and well-paid ones. The two men have discussed the nuances of this change in their books “The Second Machine Age” and “Machine Platform Crowd”.
Many cleanroom experts say we might be getting ahead of ourselves and overreacting to a sensationalised prediction. Some of the reasons they think a deserted factory will ultimately stay as science fiction are described below.
Cleaning cleanroom also throws up a lot of issues for robots. It is not as simple as using the correct disinfectant on the correct surface.
The world is too complex for an algorithm, despite technological progress, there are still tasks that only humans can do
Whether the cleaning is effective can depend on the type of wiping or what special cloth or applicator is being used. Humans can appreciate these complexities and can make judgements on if the process is working or not. For example, they can see if dust is just being whirled up. It is clear there is a risk with robots that they're doing something wrong without knowing it is going wrong. Far more advancements will need to be made before this is feasible in a machine.
Cleanroom technology is a cross-sectional technology, meaning the knowledge of over 100 professions is needed in one. From planning through construction to operation, a cleanroom is a complex undertaking. The specifications and ways of thinking from so many different disciplines under one roof, that is the challenge.
No cleanroom experts can imagine that a project as complex as hi-tech electronics production, can be managed without these specialists. The most important thing is to put together the right team, says architect Lukas Holzinger. "Who are the experts who answer the questions? How do you know you can't do this?” The robots would also have to retrofit the system and thus have the future in mind. Computer programmes are just not sophisticated enough for this yet.
The knowledge of over 100 professions is needed in one for cleanroom technology, a task that computers are not yet sophisticated enough for
Ortner says that the ability of people to decide on the spot, not only in management, but also as a project manager, assembler, or service provider, is still highly in demand. "Even if digital technology provides a wealth of data, it does not decide. The ability to judge something, to lead, to give a direction, is a human task," he says.
That's why on the planning and engineering side the most important thing for the customers is to make sure that the right people are sitting at the right table. For laboratory operator Thomas Meindl the interpersonal aspect is even more important than exchange via computer. "Projects can only succeed if there are people once sat at the table and agree. Interaction between people is crucial," he says. “In many cleanroom projects, it is important to look for creative solutions that have not yet been found. Only humans can do that."
But why can’t a machine think like a human? Even in view of rapid technological progress, there are questions that only people can, or should answer. How could a clinical robot decide how critical a patient’s situation is during surgery? This is because a robot's evaluating power is a question of its data. All computers follow a programmed routine established from input data. What is not programmed is not taken into account. Therefore any unforeseen events that are damaging may not have a pathway to follow and the response of the system will not be correct. As an example, Holzinger cites the software-supported cleanroom monitoring system’s digital limit value.
What is not programmed is not taken into account
Addressing these issues would require a 100% available, multi-cloud technology for fail-safe data storage and big data strategies for early problem identification. All this has yet to be developed and tested in a real environment, a far-off dream for today’s cleanroom users.
“Automation works well when I always do the same thing,” says laboratory operator Thomas Meindl. However, this is not the case when orders come in that require flexibility, and humans can react to the unexpected in a way machines would be unable. This scenario leads to the concept of a far more realistic machine revolution, where automation changes the workplace without replacing humans. Looking at integrating humans with technology, not replacing them, means it becomes necessary to adapt to a new production flow.
The more realistic future of integration with machines means digitalisation is not a topic of the future but of the present.
The adaptation to the new working environment has been taking place for a long time now, with mixed results. “We humans learn to deal with it quite well,” says cleanroom technician Ortner. Other experts say there are a lot of kinks that need to be worked out in order to integrate robots and people into the same working environment safely.
A few decades ago, employees in semiconductor factories had to reckon with being sacrificed in the event of a fire. When a fire was detected, the room would be immediately flooded with nitrogen to save the valuable machines, giving no time for the humans to reach safety. Today it is not customary to use the system until the employees have evacuated, but this was not the first concession made for machines, and it will not be the last.
The main reason to employ AI and machines is the belief that it will remove production errors. In reality, this belief still has its limits
Danger and emotions mean there are plenty of workers that resist any interference from machines. In the cleanroom, particle values read off by the employee are no longer deemed the height of documentation standards. Errors in manual recordkeeping have resulted in the use of computer-assisted recordings. In such an environment of mistrust, it is no wonder that the employees are afraid of the next step of digital change.
Another way in which the industry is trying to prevent machines from taking jobs is through strict SOPs, A 0% error mindset that is virtually impossible to create as there is never enough minutia to streamline every single situation, even if people could be perfect.
“When you go through the factories you can see where the behaviour doesn't work,” says Jürgen Blattner from the measuring instruments specialist, BSR Engineering Office.
The main reason to employ AI and machines is the belief that it will remove errors from production. In reality, this belief still has its limits. It is true, in the biological sense, that “electrons never carry microbes with them,” says an IT expert, confirming a statement of the laboratory technician Thomas Meindl, who is able to take advantage of this benefit of the digitalisation trend in the lab. But the sobering fact is that on the electronic level, electrons do carry issues.
Benedikt Fischer of computer consultant aConTech, explains that “we see every day what happens when non-biological intruders let off steam in systems.” He is referring to the problem of computer viruses and similar data-consuming vermin that can cause machine errors that mirror human error.
Dr Gernod Dittel
Continuing education is a natural response to job fears from technological advancement. However, people still fear for their jobs. For example, a cleaning worker at a pharmaceutical company in England feared for her job so came into work despite having the flu with a 39.5° fever. She worked in a cleanroom and when the contamination of entire batches was discovered, production was suspended for several months. When the mysterious germ source was finally ascertained, the cleaning woman was blamed for the contamination, proving fear-driven employees are a danger in the highly technological world. Better training could have made this situation less likely.
“Employees need to be empowered to adapt to change in the labour market, to cope with it,” demands the Institute for Employment Research in Nuremberg (IAB). “They need qualifications in order to take over more complex tasks, which are difficult to automate, but also to use the technologies as work equipment.”
Greater and better qualification is, therefore, a good plan, a view that is never more so valid than in the cleanroom.
The labour market researchers at the Federal Employment Agency at the IAB in Nuremberg have taken the realistic view that automation is not replacing entire professions, but instead, individual activities. Humans will likely always be necessary in some way or another, and training people to a level above the machines is the key to reducing the negative impact.
Even when humans aren’t needed, there are many studies that have questioned whether it will also be legally possible, ethically justifiable or economically sensible to employ full automation.
N.B. This article is featured in the January issue of Cleanroom Technology. Subscribe today and get your print copy!
The latest digital edition is available online.
N.B. This article has been an edited version of a feature published in German in the May 2019 issue of ReinRaumTechnik and published with permission of the author.