There’s widespread agreement that artificial intelligence — AI — is the next big thing for manufacturing. Once we harness all that data that our machinery is collecting, the thinking goes, we’ll be able to automate a lot more things than we used to be able to automate. We’ll be able to customize processes and even generate new ideas.
Automation will cost less and face fewer roadblocks. Machines will provide their own quality control and plan their own workflow. Our days at the factory will be full of the buzzin’ of the bees in the cigarette trees by the soda water fountain!
Before we get too excited, though, we should remember the old saying, “Never try to teach a pig to sing. It wastes time and annoys the pig.”
Artificial intelligence is awesome and holds many opportunities for the future. But we shouldn’t forget the limitations.
AI is currently overhyped
A computer is considered artificial intelligence if it can perform human-like tasks and show cognitive capability. Often, algorithms can capture patterns in data that people can’t see because the data set is simply too large. They can reach a level of consistent accuracy that people can’t match. They can also do a lot of things much faster than human beings.
Even more important, machines can take on the dull, dirty, and dangerous work that human beings simply don’t want to do. They can free human beings up to do things that human beings do well.
On the other hand, a lot of the AI research making headlines seems to be working on getting robots to do things that people do well…much better than robots.
This makes sense, because we are more impressed by a robot that can do something we perceive as difficult than by a robot that can move a workpiece from point A to point B. When the goal of the research is to get funding, to open people’s eyes to the possibilities of AI, or to wow the observers, it makes sense to use AI to give robots the semblance of competing with human beings.
“Semblance” is the right word here
Consider this paragraph written by AI on the subject of AI in manufacturing:
Artificial Intelligence is making waves in the manufacturing industry, with more and more companies adopting AI to improve efficiency. Abedin discusses how Artificial Intelligence is changing the face of manufacturing across all industries, including the food and beverage industry. There are so many things to do, see, eat and drink while visiting London! Here are some places you should check out while you’re there:
Who’s Abedin, what does London have to do with it, and how fast would you fire a human being who turned in that paragraph?
It’s a lot like human language, but it’s not a good job of writing about AI in manufacturing.
Boston Dynamics’ dancing robots don’t actually listen to the music and express themselves by dancing, either.
These are fun tricks…but fun tricks are entirely worthless in the factory.
The solution
It makes sense to use AI to do the things AI can currently do, and to use it to optimize automation which does the things automation does well.
The dirty, dull, and dangerous, for sure.
Also the things that require speed and precision rather than anything approaching human intelligence. Predictive maintenance, for example. Dancing and writing? Not so much.