Editor’s Note: This post is the second in a Siemens Stories series focused on artificial intelligence, “The Real Purpose of AI.” Follow along with us on Twitter using the hashtag #BehindTheAI to learn more about AI’s potential to solve real-world problems and positively impact society.
From the moment the first assembly line began running to build Henry Ford’s Model T, people have relied on machines to help mass-produce many of the products they depend on every day.
Over the last century or so, simple human-operated machines have evolved into complex robotic processes. Those robots, in turn, have allowed humans to mass-produce millions of identical items as machines have gotten better and better at repeating the same monotonous, mechanical tasks with increasing speed and efficiency.
Today, a Fourth Industrial Revolution is changing the game as, more and more, consumers demand mass customization. They want to be able to create the product they want online and receive it quickly, while still demanding top quality and an attractive price.
So, will robots be able to change with the times to meet this new era of increasingly individualized consumer demand? Will robots be able to, as the saying goes, “Learn new skills”?
Researchers at Siemens believe the answer is a resounding yes.
The traditional way of using robots in manufacturing has revolved around programming robots to do one specific task and then repeat it again and again. Robots were isolated from humans through fences. And programming was a tedious and costly process. That production model served industries well in an era of mass production.
But as the Fourth Industrial Revolution progresses and mass customization becomes a reality, humans who run machines now have to create more complex programming in order to have a higher level of autonomy. And if the same machine is required to do another job, it has to be reprogrammed from scratch, which is no small task.
This is where artificial intelligence comes in. Rather than create increasingly extensive computer programming, companies like Siemens are rethinking automation and developing self-learning robots that can learn from their experience in much the same way the human brain does. Additionally, companies are rethinking autonomous engineering systems that enable a high-level programming of skills instead of instructions.
Instead of telling a robot exactly what to do, programmers tell the robot what task needs to be achieved. The robot can then “learn” through a process of trial and error how to use its tools to accomplish that task. Or when to use simulations or knowledge graphs to speed up the process. However, AI alone won’t provide the necessary reliability. We need a combination of automation, digitalization and AI to advance the era of autonomous systems in manufacturing. We need flexible yet robust algorithms running on industrial grade hardware.
In the case of an automobile manufacturing line, we can envision an autonomous application where a robot picks out a 3D-printed part that was never seen before and uses the tools at its disposal to install the desired part before moving on to the next different part. The end result is an AI-powered autonomous robot that can adjust its strategies and handle increasingly complex work and increasingly complex dexterity and decision-making.
While the applications for autonomous robots in the industrial sector seem limitless, there remains some concern about widespread automation replacing humans. In fact, the more likely future is humans working alongside robots. As Siemens USA CEO Barbara Humpton wrote, AI’s purpose isn’t to automate humans out of the process but to expand what is humanly possible. “And that’s because, yes, AI can help us diagnose, it can help us treat, it can help us understand – but it takes the human in the loop to truly develop the solutions that we’ll need for the future.”
As robots learn new skills, humans will learn new value-added tasks. As they acquire that knowledge, they also have more time to innovate and shape the future of the industrial world.
Published On: April 10th, 2019