BALTIMORE – While economic pundits and policymakers opine on the decoupling of the US and Chinese economies, a revolution that will bring the opposite is already underway.
It’s increasingly clear that industrial applications of artificial intelligence will make economies more global, despite geopolitical tensions that grab much of the world’s attention.
In fact, as an Asia Times webinar audience discovered Tuesday, the dynamics driving industrial innovation make closer integration more critical than ever before.
The webinar brought together panelists intimately familiar with the factory floors and corporate offices that are the driver’s seat of a 4th Industrial Revolution transforming business at a dizzying pace.
The linchpin of this transformation is the collection and analysis of large swaths of data to create AI algorithms making industrial processes more efficient.
Maximillian Dommermuth of German engineering giant Bosch Rexroth AG underscored that any players dragging their feet will get left in the dust. Dommermuth, head of training, digital transformation at Bosch, illustrated the dynamic at play:
“Take one machine, and record data for twenty thousand days. You would need a little bit more than fifty years and you get your data set, and then you can produce a nice AI algorithm… Or you can take twenty thousand of the same machines around the world and collect their data. Then you need one day to do the same AI algorithm,” he said.
“Technology implementation,” Dommermuth stressed, “goes far beyond purchasing technology.”
“When you buy a car, you don’t purchase a new factory. Similarly, when you invest in technology,” he explained, “you’re not acquiring an entirely new organization. Instead, it’s like fine-tuning your car, upgrading it with new technology, and finding yourself in a modern cockpit…
“You have dashboards, visualizations, and all the necessary technology for autonomous driving. However, when you try to accelerate, you encounter an unexpected roadblock. The car doesn’t move, and your company remains stagnant. You realize that something crucial is missing — the tires, symbolizing the lack of IT infrastructure required for progress.
“Additionally, you discover another missing element — fuel, representing the absence of necessary data for important tasks. As you inspect further, you find that everything appears to be rusted through, highlighting that you lack the right people and qualifications to kickstart the initiatives.”
Former CTO of China’s Huawei, Paul Scanlan, who is currently an advisor to the telecom giant’s president, was emphatic that to benefit from AI, global collaboration is crucial.
“The only way you’re going to be able to scale any of the solutions that we’re talking about, whether they’re focused on circularity and sustainability or focused purely on efficiency or innovation is to collaborate,” Scanlan said in his webinar remarks.
Technological improvements will eventually be adopted universally, he added, as they tend to be once they’ve been proven successful. “At the end of the day, one of us is going to look at the other and say, ‘that’s pretty brilliant’ despite which country we come from or which company we are working for.”
Dommermuth explained that this plays out across industry sectors and company sizes.
“When it comes to aspects such as infrastructure, strategy, and scaling up processes, there are commonalities. It’s not limited to mass production; the challenges are unique in their own way. Mass production, as exemplified by Taylorism during past industrial revolutions, involved producing the same product repeatedly to boost productivity and efficiency. But today, customers dictate what they want to buy. As a result, manufacturers have to adapt to smaller lot sizes,” which Dommermuth says places a greater reliance on some of these new technologies.
“As you move towards smaller lot sizes, the importance of having a robust data infrastructure becomes even more evident. It enables the automation of various tasks that would otherwise be labor-intensive if done manually.”
Huawei’s much-touted Pangu system has attracted attention for its industry-specific AI models, which Scanlan says have already been deployed across a wide range of industries.
“We have models that do everything from mining and manufacturing to drugs. Just take drugs for instance, we’ve trained this system to decode 100 million naturally occurring compounds. Then it only takes about six weeks to make a match between a disease and drug. And by the way, that’s not really a drug. It’s a naturally occurring compound. So suddenly the cost of the drug is 70% to 80% less.”
“Take freight trains,” Scanlan continued, “we take millions of photographs of the rolling stock, watch the behavior, train the AI and then maintenance costs fall significantly. And that translates directly to a conveyor belt in a manufacturing facility or a mining plant, or a food production plant.”
Ultimately, the panelists agreed, technological innovation will be pushed by industry needs not geopolitical trends.
“If your targets are similar, you should work together because you can learn from each other and you can get the right application and the right way to drive this… if you look at the world, we all want to be efficient, we will all want to be driving our new developments,” Dommermuth said.
“If you don’t start today,” he cautioned, “You cannot accelerate later because you won’t get on the road. It means you have to start now so that you can really get your benefits out of it in the end.”
If you missed it, you can catch the full webinar, 4th Industrial Revolution: De-Risking or Re-Coupling?, on Youtube by clicking here.