Pinsetter. Telegraphist. Typist. Those jobs are gone, obsoleted by advances in technology. But where are the marauding hordes of unemployed pinsetters? Well there aren’t any. They all went on to something else. Maybe they’re now web designers or Instagram Influencers. The point is, technology takes away some jobs and creates new ones.
A lot of people in manufacturing are concerned about Artificial Intelligence (AI) and robotics taking away jobs. It was even a topic of discussion at the recent FabTech show in Chicago. And the consensus among people who actually understand what this technology is all about? Don’t worry about it. Yes, changes are coming, but they’ll create opportunities and growth. That sounds good, but we wanted to know more. Here’s what we learned.
Robotics in Manufacturing
You’re probably as familiar with industrial robots as we are. While they take many forms it’s the five or six axis articulated arm machines that are most often seen at work in manufacturing. Welding and paint spraying, especially in automotive assembly plants, were some of the first jobs taken over my robots. More recently they’ve added palletizing and depalletizing, assembly, packing, deburring and machine tending to their resume.
Robots excel at performing repetitive motions. They don’t get tired and their attention doesn’t wander. That makes them ideal for what are often described as dirty, dangerous and dull jobs. They do however have a couple of downsides. First, they need some setup and programming, and this makes them impractical for short runs. And second, they can’t cope with unexpected variability, like when a welder finds the blanks he’s welding are cut to slightly different sizes.
Collaborative robots, or cobots for short, are the next big thing in manufacturing. We don’t have any of these on-site just yet: ours are the typical 5/6 axis articulated arm machines, but we are looking at cobots.
Cobots are engineered to be safe. Typically there’s some kind of power and force limiting technology that prevents them from striking a human hard enough to cause injury. That does away with expensive safety cages, although a risk assessment is still needed. (Even if the robot is safe, what about the grinding wheel or welding torch it’s carrying?)
Cobots are also easy to teach. For most it’s just a case of guiding the gripper through the path you want it to follow. That makes setup quicker, so a cobot can, in theory anyway, be used on shorter runs than a conventional robot. Many manufacturers are exploring the potential of cobots for jobs like machine tending, and it seems safe to say we’ll see growing numbers in factories in the the years ahead.
AI and Manufacturing
Unlike robots, which have been around a long time, AI is only just emerging as a tool for use in manufacturing. There’s a lot of jargon and buzzwords around AI, which we’ll try to cut through.
AI is basically software that can learn to recognize patterns. Based on what it learns, it can sift through huge datasets to make recommendations.
The key to making AI work is to have lots of data about manufacturing processes, which is where the Industrial Internet of Things (IIoT) comes in. The ideas is that sensors have become really cheap, and it’s easy to connect these sensors to extract the data from them. (There’s Ethernet, WiFi, Bluetooth, MTConnect, OPC UA and so on.) So lots of machines can send data about what they’re doing back to a central point. (Speeds, feeds, vibration, temperature, humidity and power draw are just some of the examples.)
At that central point analytics software starts doing the sorting and analyzing. Over time it learns to identify trends and patterns and can raise warnings when it thinks something will need attention or suggest ways to improve yields, utilization and similar metrics.
Manufacturing Applications of AI
Two areas where AI is seen as having real promise are:
- Predictive maintenance
- Process control
In predictive maintenance applications AI can identify when a machine of piece of process equipment is likely to need attention. It could log vibration to sense bearing wear, or take data from multiple sensors to warn when filters will need replacing. This will reduce the need for time-based maintenance, increasing availability and capacity.
In process control AI could monitor raw material conditions and adjust a process to suit. For example, if material humidity is above average, increase time in the oven. It could also monitor in-process parameters like fluid temperatures and motor torques to determine when a reaction is complete. In complex plant environments AI could optimize operating conditions to reduce waste and lower energy costs.
The thing about AI in manufacturing is this: it’s new and we’re only just starting to think what it could do.
Implications for Jobs
Let’s start with a reality check: manufacturing is facing a skills gap. The workforce is getting older and fewer young people are entering fields like fabrication and welding. This was discussed in a recent story on CNBC.com about a study by Deloitte and The Manufacturing Institute. This projects that 4.6 million manufacturing positions will open up over the next ten years. The problem is that only around 2 million of those positions will be filled.
Unless we can get more productivity from the people we have a lot of manufacturers are going to struggle. This is where robots and AI come in. These technologies aren’t going to replace jobs, they’re going to leverage the skills of the people who do those jobs now so they can do more.
Welding offers a good example. Fabrication requires a lot of welding work, and we have both manual and robotic welding systems. Welding is a highly skilled job, perhaps almost as much art as science, and good welders don’t grow on trees. But what if those welders could set up and oversee robotic welding systems? Maybe in the future some of those systems will use easy-to-teach cobots, and perhaps there’ll be AI tools to help optimize welding parameters on each job. This way each human welder will be responsible for more output, perhaps while being kept away from fumes and arc flash.
The details are unclear right now, but we know we will need machines to handle those dirty, dangerous and dull jobs because there won’t be enough people to do the work. Plus, having freed-up workers from repetitive tasks we can ask them to do more problem-solving and process improvement work: things that need doing but we don’t have the time right now because of all the routine tasks that get in the way.
Rising Productivity Grows Our Economy
Maybe this all sounds like a lot of pie-in-the-sky utopian nonsense, but really it’s just a continuation of what’s been happening since the Industrial Revolution. New machines take on work previously done by hand, increasing productivity, lowering costs and letting people develop and use new skills. You might say it’s the American Way of progress.
Let’s finish by talking about one of those obsolete jobs: that of pinsetter. You once found pinsetters not in garment manufacturing but in bowling alleys. These were the people who stood the pins up ready for the next bowling ball. Today that work is done by a machine, and probably faster and more accurately than a human could do it. And what of the pinsetters? Maybe they went back to school and became engineers. We’d like to imagine a different career for them though. Perhaps, having seen so many bad bowlers, they all became bowling instructors.