China Development Research Foundation   |  APP Download   |   中文   |   Register   |   Login
Time:March 23-25, 2019
Beijing Diaoyutai State Guesthouse
Sponsor:Development Research Center of the State Council
Organiser:China Development Research Foundation
CDF WeChat
CDRF Weibo
Background Paper Latest from CDF
Back to Background Paper List>

2018丨【ABB】Manufacturing in the age of artificial intelligence

Executive Summary

The era of artificial intelligence-enabled manufacturing has arrived. While “artificial general intelligence” (a machine that could perform any intellectual task as well as a human) or “conscious AI” (a machine with a software-based consciousness) remain far out, we have already created and deployed “early industrial AI.” This type of AI is generally comprised of software that automates or augments an activity formerly performed by humans. Machine learning, which refers to the capacity of AI to rewrite its own software in response to new data, is currently receiving most of the funding dedicated to the development of AI. Such systems are starting to appear on the factory floor, augmenting the capacities of industrial automation systems, which many believe already comprise a form of early AI.

Worldwide, investment in artificial intelligence is expanding rapidly. Progressively sophisticated algorithms and more powerful computers, together with the proliferation of data and affordability of data storage, have increased the likelihood that we will witness a step change in the capacity of AI systems in the near future. While “digital natives” such as Baidu and Google are pushing ahead, other sectors have lagged behind, uncertain about how to procure, integrate and deploy this complex new technology. Despite the promise of AI-powered production, manufacturers have still to develop clear methods to take these technologies from proof-of-concept to industrial scale. While many challenges remain, the potential benefits of AI for the manufacturing sector are enormous.

In 1974, ABB created the first commercial, electromechanical robots operated by software. Today, ABB is leading in the capabilities of the industrial internet and exploring the potential of AI in industrial production. To stay in front in the field of automated production, ABB is setting the pace with emerging innovations in the fields of machine learning, smart structures, augmented reality and autonomous mobility. The company is starting to see concrete results from an agreement that it signed with IBM in 2017 to develop the machine-learning capabilities of ABB AbilityTM, its industrial digital offering. By pairing ABB Ability with IBM’s Watson computing system, the objective is to move beyond current connected systems that simply gather data into a new era of cognitive industrial operations. AI excels at interpreting large amounts of data and applying algorithmic-based solutions, but lacks human intuition and creativity; ABB aims to use AI’s strengths to supplement and augment the strengths of human experts.

Over the past few years, the Chinese government has clearly demonstrated its commitment to comprehensively upgrade the country’s industrial base through the deployment of advanced automation and AI technologies. In China, AI and other disruptive technologies have, to date, primarily been deployed by the consumer goods sector. The widespread application of AI to the industrial and B2B sectors, however, will be a more critical test of the technology’s potential to reshape the economy. Importantly, AI-fueled productivity gains will be paired with massive changes in China’s labor market. Research studies estimate that some 51 percent of all jobs in China could be automated, more than in any other country in the world. Yet it would be a mistake to interpret this forecast to mean that AI will lead to large-scale job losses. A far more likely scenario is that by increasing productivity and therefore competitiveness and prosperity, AI will lead to new jobs and ultimately new industries which will create many more jobs than will disappear as a result of this new technology.

AI cannot supercharge a factory’s productivity on its own. First, there must be a concerted effort to fully digitize every aspect of a company’s operations, networking them together to create a “smart factory” that generates and transmits high-quality structured data about every aspect of its value chain. A smart factory, by definition, is able to optimize its own manufacturing processes automatically. By providing such a smart factory with an AI system, the full potential of this new technology can be realized. Today, many firms are laying the foundations for smart, AI-powered factories by installing connected smart sensors and using powerful, cloud-based algorithms to boost uptime, speed and yield. Foxconn provides an example of how one major multinational manufacturer is preparing its shop floors for a series of AI breakthroughs.

While intelligent manufacturing is still in its early stages, new modes of production will ultimately be capable of autonomous learning, independent decision-making, self-modification and even self-perception. The potential impact of AI on intelligent manufacturing can be discussed in terms of three key fields of technological development – intelligent equipment, intelligent factories and logistics, and intelligent services and maintenance. A sequence of short case studies are used to demonstrate how ABB is deploying these technologies in industries across China, as well as at its own sites around the world.

AI has the potential to reshape China’s manufacturing sector in the years to come, yet the right conditions must be created for this transformation to take hold. This paper recommends that China: 1) strengthen the “Industry 4.0” aspects of its “Made in China 2025” initiative; 2) create strong, cross-sector rules to ensure all the data produced in China is machine-readable; 3) drive the adoption of AI technologies across government; and 4) recalibrate the educational system to create a large pool of skilled workers and encourage the development of uniquely human capabilities that cannot be automated.