US factories at forefront of artificial intelligence adoption
Though most industrial firms around the globe are eager to implement AI tech into their processes, a much smaller amount have already integrated AI into their businesses. US firms, are however, at the forefront of early adoption, with a quarter of companies reporting that they have already implemented more than one AI use case.
Artificial intelligence is set to revolutionize pretty much every industry, as machines process huge amounts of data and ‘learn’ more efficient ways of performing processes. One area set to adopt the technology perhaps more quickly than others is the manufacturing sector, which previously embraced the functionality of non-AI robotics to improve productivity. Using machine learning algorithms – which learn from the analysis of large data sets and then generate predictions – factories can boost efficiency, accelerate processes, and enable ‘self-optimization’ operations. Through the use of AI, producers can lower conversions costs by 20%, according to The Boston Consulting Group.
In order to better understand the state of AI adoption in the industrial sector, BCG conducted a survey of more than a thousand executives and managers from producing industries, titled ‘AI in the Factory of the Future: The Ghost in the Machine.’ The survey revealed that firms in US, China, and India have taken a strong lead over other countries like Japan, Germany, and France in the early adoption of AI in production processes.
AI will augment the continual improvement processes that industrial firms use to accelerate productivity, like lean management. For example, AI can use machine vision (x-ray, laser signals) to examine parts and products, and then learn from data the cause of defects and reduce production waste – in line with the goals of lean management. Indeed, 40% of BCG survey respondents said they expect AI to be a very important driver of productivity improvement in 2030, while 29% believe it is very important to productivity improvement today.
One unavoidable aspect of AI introduction into production processes is the reduced need for manual labour. Most respondents believe that the net effect of AI will be a reduction of the workforce – the jobs created to complement AI processes will simply not outnumber the number of jobs lost. However, expectations vary from country to country: respondents from China believe that AI adoption will significantly reduce their total workforce, while German companies expect smaller reductions to their more highly-skilled workforce.
Today’s factories automated processes follow a rules-based approach that conforms to a fixed scenario. AI, on the other hand, will allow machinery to respond to unfamiliar situations with ‘smart’ decisions. For example, AI-enhanced machinery will be able to identify and select parts from an unsorted bin, while a rules-based robot would fail to cope with the many possible orientations of the parts.
37% of respondents rated production as the area where AI would be most important avenue to increase productivity, while 25% said quality would be most important, and 12% picked logistics. Likewise, respondents believe that self-optimizing machines, defect detection, and prediction of efficiency losses are the most important AI use cases in the industrial setting.
Inside the factory, AI integration will spur improvement to production, maintenance, quality, and logistics processes. In production, AI will enable machines to become self-optimized systems that adjust parameters through analyzing and learning from the data they gather. For example, steel mill furnaces are already using AI to self-optimize their settings: AI analyzes the iron intake and identifies the lowest temperature needed for stable processing – reducing overall energy consumption.
AI will also improve the maintenance process within factories, avoiding breakdowns by replacing parts before they actually break. Analyzing and learning from sensor data and product data, AI will reduce downtime due to equipment failure. Oil refineries have already installed machine learning models that more accurately determines equipment failure by considering more than 1,000 variables – from material input to weather conditions. Likewise, AI can use vision systems to identify defects in products: automotive suppliers have started using machine learning and vision systems to identify defects, including ones not used in the data set training the algorithm.
In-factory logistics and warehousing is already being revolutionized by AI-enabled autonomous movement. Self-driving vehicles can efficiently move supply, learning the best path and avoiding obstacles – indeed, health care equipment producers are already using self-driving vehicles in their repair centers.
Over 80% of survey respondents expect the above use case to be very important by 2030; the belief that the capabilities have already been fully implemented in multiple production areas, is unsurprisingly, quite low – ranging from 6-8% between various use cases.
Firms in most countries are eager to implement AI in the near future – though China, India, and Singapore are the most eager – with 94%, 96%, and 97%, respectively, responding they plan to implement AI within the next three years. 87% of US respondents had ambitions for AI integration in the next three years – which was also the global average.
However, only about 16% of global respondents have fully implemented more than one AI use case in multiple plant areas – which BCG classifies as ‘early adopters.’ Indeed the US has the highest percentage of early adopters, with 25%, while China and India rank second and third with 23% and 19%, respectively. The countries with the smallest proportion of early adopters were Japan (11%), Singapore (10%), and France (10%).
The high level of adoption among US industrial firms reflects the widespread availability of AI technology in the country, and the presence of many firms offering AI tech and implementation. China, however, has overtaken the US in AI funding – with healthy support from government authorities, including the recent announcement of $5 billion in AI funding from the Tianjin municipal government. Likewise, India regards AI adoption as essential to keeping its producing industries globally competitive, and has also made large investments. In contrast, some industrialized countries, like Japan, remain more focused on more conventional productivity levers like lean manufacturing, which promoted its competitiveness in the past.
Among the producing industries surveyed, health care and energy are the most eager to implement AI in the near future, while process industries and engineered products industry respondents are less so. Among the eight industries, transportation and logistics (21%) and automotive (20%) have the highest proportion of firms that are early adopters, while engineered products (15%) and process industries (13%) have the smallest proportion. BCG notes that transport, automotive, and tech firms are the most advanced in terms of AI adoption because digitization, digital tactics, and/or automation have already been integral parts of their value chain for years.
