“I think we May exceed the one-to-one ratio of humanoid robots to humans,” Elon Musk declared on March 1. The self-proclaimed technology from Tesla is less a prediction than a promise. Mr Musk, whose car company is developing an artificial intelligence autonomous machine code-named Optimus for use in homes and factories, made the remarks during Tesla’s investor day, accompanied by a video of Optimus apparently walking independently .
Given that Mr. Musk didn’t detail how — or when — you’d get an army of more than 8 billion robots from the promo, this might smack of science fiction. But he has stepped into a very real debate about the future of work.for some forms of artificial intelligenceSupported automation is fast becoming a science fact.
chat since novembercommon technologyone artificial intelligence Conversationalists, which dazzle users with their general impressions of human interlocutors. Other “generated” artificial intelligenceHuman-like texts, images, and voices have been imagined by analyzing vast amounts of data on the internet.boss last month IBMa computing giant, predicts artificial intelligence A lot of white collar paperwork will be eliminated. On March 6, Microsoft announced the launch of a set of artificial intelligence A “co-pilot” for workers in jobs ranging from sales and marketing to supply chain management. Excited observers grumble about the impending working apocalypse.
Of course, concerns about the impact of technology replacing jobs are nothing new. In England in the early 19th century, Luddites burned factory machines. The term “automation” first came into focus in the 1950s, when the adoption of wartime mechanization innovations sparked fears of mass unemployment (see Figure 1). In 1978, British Prime Minister James Callaghan welcomed the breakthrough technology of his era – the microprocessor – and the government investigated whether it was killing jobs. A decade ago, Carl Frey and Michael Osborne of the University of Oxford published a blockbuster paper, cited more than 5,000 times, claiming that among tasks performed by American workers 47% could be automated “within the next ten or two years.” Now, even Mr. Musk, a techno-optimist, wonders what it means for robots to outnumber humans: “It wasn’t even clear what an economy was.”
Although Mr. Frey and Mr. Osborne still had years to be proven right, and Mr. Musk could be safely ignored for the time being, earlier fears of job-killing technology never materialized. In contrast, labor markets in developed countries have historically been tight—and are becoming structurally tighter as societies age. There are currently two job openings for every unemployed American, a record high. The US manufacturing and hospitality sectors reported labor shortages (measured by the gap between job vacancies and unemployed workers whose last job was in the relevant sectors) of 500,000 and 800,000, respectively.
Fury of the Machine
The immediate problem in advanced economies, then, is not too much automation, but too little. Worse, for large enterprises, automation is often difficult to get right in practice.The latest buzz may well prove it won’t be easy artificial intelligenceSecond.
Robotic arms that perform repetitive tasks like welding, drilling or moving objects on factory floors have been around for decades. The use of robots has historically been concentrated in the automotive industry, where heavy parts and high volumes with limited variety are ideal for machines. The electronics industry, which requires precise but repetitive motion, was also an early adopter.
Jeff Burnstein, president of the Association for the Advancement of Automation, a US trade group, pointed to a recent increase in the list of industries adopting robots.Advances in computer vision are making machines more dexterous, says Sami Atiya, who runs the robotics business ABB company, a Swiss industrial company. Lightweight “collaborative robots” now work side by side with human workers rather than being kept in cages, and self-driving cars transport objects from one place to another in factories and warehouses.
At the same time, robot prices plummeted. Asset management firm Ark Invest estimates that the average price of an industrial robot fell from $69,000 in 2005 to $27,000 in 2017.in december ABB company Opened a 67,000-square-meter “megafactory” in Shanghai where robots build other robots. Installation costs have also come down, and new “no-code” systems require no programming skills, notes Susanne Bieller, secretary general of the International Federation of Robotics (international flight rules), another industry body.
Thanks to better technology and lower prices, the global stock of industrial robots has grown from 1 million units in 2011 to nearly 3.5 million units in 2021 (see Figure 2). Sales at Fanuc, Japan’s big robotics maker, rose 17% last quarter from a year earlier; shares in Keyence, a Japanese firm that provides automation consultants for factories around the world, soared 24%. While they have retreated from their bubble peak in 2021, when chief executives were looking for alternatives to a workforce incapacitated by covid-19, robot makers are still a fifth higher than they were before the pandemic (See Figure 3).
However, despite all this growth, the absolute level of adoption remains low, especially in the West.according to international flight rules, even the South Korean company, by far the world’s most enthusiastic adopter of robots, employs ten manufacturing workers for each industrial robot — a far cry from Mr. Musk’s vision. In the US, China, Europe and Japan, the figure is 25-40 to 1.According to the consulting firm, the $25 billion BCG, global spending on industrial robots in 2020 is less than 1% of global capital spending (excluding energy and mining sectors). People are spending more on sex toys.
Rainer Brehm, head of the factory automation division at German industrial giant Siemens, has observed that the long lifespan of industrial equipment limits how quickly older, dumber machines can be replaced by new, smarter ones. Today, most menial jobs in advanced economies are anyway in service industries, where manual labor is more difficult to automate (see Figure 4). Offering 244 planes of motion, the human body’s joints and fingers are a marvel of versatility. A typical robot has six such “degrees of freedom,” says Kim Povlsen, CEO of Universal Robots, a maker of industrial robot arms.
The automation of office work has similarly stalled for similar reasons of legacy systems and corporate inertia. In theory, digitization should remove much of the human involvement from mundane tasks like ordering inventory, paying suppliers or checking out.
In fact, most businesses born before the digital age use a bunch of outdated and incompatible systems, notes Cathy Tornbohm of research firm Gartner.instead of pocket it As consultants come to clear the jungle, many companies prefer to outsource menial office work to low-cost countries such as India or the Philippines. data centeranother research firm, The $20 billion a year market for software that automates mundane office tasks is even less than what is spent on physical robots.
Automatically for the people
Over time, further innovations may remove some of these barriers. For physical robots, this is well underway in machine-crazy South Korea. Doosan Robotics, one of South Korea’s largest robot makers, has opened up its software to outside developers to create pre-programmed applications for its robots. These are now used in everything from making coffee to laying floors on construction sites. Robert Chicken uses a robotic arm to operate fryers at its fast-food restaurants; to reduce the upfront investment for franchisees, the company rents the robots to franchisees for around $900 a month, significantly less than human operators. South Korean internet giant Naver has a division that develops robotic vehicles that can navigate busy environments with complex layouts: armies of such robots are already running around delivering lunch boxes and packages to employees.
Office process automation is also becoming increasingly complex. UiPath, a pioneer in automating mindless tasks such as copying and pasting information from one program to another, now offers other tools that use image recognition algorithms to extract data from paperwork or Map business processes. Rob Enslin, Partner at UiPathCEO, said the company has served 10,000 customers. Power Automate, a Microsoft tool that allows the average desk clerk with little programming experience to automate tasks like expense or travel approvals, now has 7 million monthly active users, says Charles, who oversees many of the software giant’s automation products. Said Charles Lamanna.
Some businesses are tentatively starting to embrace generative artificial intelligence, also. However, as with robotics and process automation, the adoption of new technologies will not happen overnight. Allen & Overy, a law firm launched a virtual legal assistant with Chat in Februarycommon technology— like Powers, requiring its lawyers to cross-check everything the bots spit out. central networka tech news site that has quietly posted 73 bot-written articles since November, first stunned and then delighted reporters after discovering the articles were riddled with errors.
this artificial intelligence Mr Lamanna believes that the technology underpinning chatbots could one day be a boon for automation. But going from science fiction to science fact is one thing. Deriving economic facts from there is another matter. ■
To stay on top of the most important news in business and technology, subscribe to The Bottom Line, our weekly newsletter for subscribers.