Surprisingly, most workers today are optimistic about artificial intelligence (AI). Amid doomsday headlines about massive job losses due to automation, more than half of employees would welcome AI if it reduces time-consuming drudge work (64%), increases fairness (62%), or ensures that managers make better choices in the workplace (57%), according to a 2018 Workforce Institute survey of 3,000 workers across eight industrialized nations.
But workers are anxious about AI, too—and that anxiety is directly linked to a lack of communication from company leaders, the survey finds.
The problem: Only 38% of firms have a clearly defined AI strategy at present, according to the Boston Consulting Group. As companies evolve their digital transformation plans, a lack of transparency from the C-suite is understandable, but is it ethical? Experts believe that companies owe some measure of transparency to their workers, and beyond mere words, some feel they have an ethical obligation to take action—namely, to equip their workers with new skills for the coming age of AI.
“The No. 1 source of worker anxiety is around ensuring the relevance of their skills,” says H. James Wilson, managing director of consulting firm Accenture and co-author of “Human + Machine: Reimagining Work in the Age of AI.” “Reducing this concern is the ethically responsible thing to do.”
But three out of every five organizations globally (58%) have yet to discuss the potential impact of AI on their workforce with employees, the Workforce survey finds. U.S. companies are the most secretive: 67% of U.S. workers surveyed say they have no knowledge whatsoever about their organization’s plans for AI.
Moreover, data also shows that the introduction of AI automation into the workplace can negatively impact employees’ well-being. A 10% increase in automation worsens workers’ physical and mental health, lowering their overall health by 2.38%, a study by Villanova University and Ball State University finds. On the surface, the number may seem low, but researchers say it has a compound effect: A 20% rise in automation will double the health impact, 30% will triple it, and so on.
From an ethical standpoint, what can business leaders say to put their stressed-out workers at ease? As the issues at play come into focus, forward-leaning companies are taking the right steps through both words and action.
What Companies Can—And Should—Say
Almost all executives know there’s an AI anxiety problem; 87% of executives surveyed recently by ServiceNow say their employees are worried that automation will eliminate jobs. But they feel that those fears are misplaced: 79% of those surveyed believe that AI will lead to job creation, not losses.
Some research supports this sunny view. Within two years, AI could create 2.3 million more jobs to offset the 1.8 million it may replace, according to a Gartner study. Even sunnier, as many as 58 million new jobs could be created by AI by 2022, according to a 2018 World Economic Forum report.
But other research is less optimistic. By 2030, AI automation could eliminate jobs for as many as 30% of the world’s labor force—between 400 million and 800 million jobs—according to a two-year study by McKinsey, which predicts that AI-fueled economic disruption could rival the massive upheaval brought by the shift from agricultural to industrial labor during the early 1900s.
With each new study inspiring whiplash-inducing headlines, what can workers believe? The truth, most likely, is neither utopian nor dystopian, according to Robert Seamans, associate professor of management at NYU’s Stern School of Business.
“There’s a lot of misinformation out there,” Seamans says. “The narrative that AI is coming to take your jobs is way overplayed.”
Conflicting studies notwithstanding, companies are investing heavily in AI. Enterprise investment reached $19.1 billion globally in 2018, a 54% year-over-year increase, according to an IDC report. And corporate AI spending will hit $52.2 billion by 2021, the consulting firm projects.
Given this rapid escalation in AI investment, saying nothing to employees is almost certainly a worse ethical choice for business leaders than saying something—anything—even if it means admitting they can’t say much yet, argues Thomas Davenport, distinguished professor at Babson College and author of “Only Humans Need Apply: Winners and Losers in the Age of Smart Machines.”
“Telling employees about potential AI job impact may not be an ethical responsibility, but it is good management,” notes Davenport. “It allows workers to start retraining and preparing themselves to adopt new roles working alongside AI-based systems. And it facilitates their collaboration in developing and implementing the new systems.”
That said, companies must walk a fine line in crafting an honest message, without overpromising that no worker’s job may be at risk.
“They could say, ‘We’re not sure exactly what will happen with AI, but we expect a high degree of AI and employee ‘augmentation’ rather than large-scale automation,’” Davenport suggests. “‘And we will try to make workforce reductions through attrition whenever possible.’”
The AI-Augmented Worker
Augmentation or automation? Communicating the distinction to workers could play a huge role in reducing AI anxiety, according to Seamans.
“It’s really important for businesses to get the message out to their employees that these technologies will augment our existing jobs,” he explains. “One way to do that is to tell stories about the different ways that firms are using these new technologies to complement what their employees are doing.”
Among current examples of AI augmentation that business leaders can cite? Workers at more than 50 major manufacturers such as Boeing, DHL, GE, and Volkswagen are using AI-enhanced smart glasses to guide repairs and inspections. Warehouse staffers at Amazon and Walmart, as well as assembly line workers at Tesla, Ford, and GM, are using cobots to automate repetitive processes and eliminate heavy lifting on the factory floor. Major financial institutions and insurers are training algorithms to detect fraud, reducing millions of work hours spent by analysts on mundane processing.
Experts say AI augmentation across industries is both inevitable and profitable. By 2022, one in five workers will collaborate with an AI “co-worker,” a Gartner study predicts. And only a small sliver of occupations—less than 5%—are candidates for total automation, according to McKinsey. But a full 45% of workplace tasks could be automated by AI and robots, the research firm finds. The savings to companies in wages could amount to nearly $15 trillion.
Accenture’s Wilson explains that companies must convey to workers that automation of some work processes will improve their jobs, not eliminate them. “When you announce or imply that a new AI initiative is all about ‘automation,’ and not about human and machine collaboration, you can lose the buy-in and support of your best people,” he cautions.
Research from “Human + Machine” bears this out, Wilson notes. Companies that focus solely on automation see modest productivity gains over the short run, but those gains soon fizzle. However, businesses that emphasize human and AI collaboration significantly outperform “automaters” in speed, scalability, flexibility, and decision-making effectiveness.
“The greater the degree of organizational focus on people helping AI, and AI helping people, the greater the value achieved,” Wilson says.
The Duty To Reskill
As AI becomes more prevalent, two-thirds of executives (67%) recognize a growing skills gap in their company, according to an Accenture survey. And only 26% of senior executives believe their people are ready to work with intelligent technologies.
Expect the skills challenge to only increase. By 2030, AI automation may require as many as 375 million people to not only master new skills but to switch job categories entirely, McKinsey predicts.
For Josh Bersin, human resources expert and founder of Bersin by Deloitte, the question of whether to reskill current workers or hire new talent is a no-brainer.
“Helping people reinvent themselves may be the biggest key to achieving business and social prosperity in coming years,” Bersin says. “Making this commitment is proof that a company cares about their employees and values their contributions.”
He points out that reskilling ultimately saves money. “Companies that invest in reskilling employees report increased employee retention,” notes Bersin. “Given that one out of three new hires typically doesn’t work out, training current employees for new roles makes sound business sense.”
Despite this, many companies are resistant to devote resources to reskilling. An Accenture survey of 1,200 CEOs uncovered a disconcerting disconnect: While 74% of business leaders plan to use AI to automate tasks to a large or very large extent over the next three years, only 3% intend to increase investment in worker reskilling over the same time period.
“Surveys I’ve worked on suggest that more than half of U.S. executives in large companies don’t feel they have an ethical responsibility to reskill their workers,” Davenport says bluntly. “They don’t plan to reskill—they plan to hire new people.”
Davenport predicts the strategy will prove shortsighted. “This disturbing news will probably become apparent to employees, who will then begin to withdraw their collaboration and their expertise in helping to implement intelligent systems,” he says.
But Seamans disagrees—because many companies stand to benefit by retaining top talent.
“Frankly, it’s more cost-effective to reskill an existing workforce than to fire a segment of the workforce and hire a new set of workers with different skills,” Seamans argues.
Several major companies agree. AT&T has invested $1 billion in a sweeping program to retrain its workforce with digital skills. Through its IBM Skills Academy, more than 200,000 IBM workers have acquired advanced AI skills. And Bank of America has launched a reskilling program geared toward workers whose jobs may be at risk due to automation.
Retail giant Walmart is also tackling reskilling on multiple fronts. Since 2015, it has invested $2.7 billion in training programs and raising wages. Its Walmart Academy has graduated more than 200,000 associates and managers, and its Pathways program brings new hires up to speed on computer skills.
Intriguingly, it has brought cobots into stores to “train” retail workers on how to collaborate with them. The bots have taken over mundane shelf-scanning and floor-cleaning duties, freeing human employees to spend more time assisting customers. Anecdotal evidence in the MIT Technology Review suggests that human staffers are championing their robot coworkers to customers, even giving them names.
Accenture’s Wilson thinks Walmart’s program represents an ethical approach to introducing AI augmentation in the workplace. “Walmart showed the importance of a ‘getting to know you’ period,” he explains. “A worker may feel it’s an ethical violation if they don’t get a proper introduction to their new AI colleague. Doing a week- or month-long demo helps workers understand how the AI works, which tasks it will handle, and so forth.”
“Leading companies like Walmart are showing the importance of small-scale experimentation so employees feel engaged in co-designing the AI initiative,” he adds.
From his perspective, Seamans sees the ethical issues raised by employee anxiety as part of a continuum, as enterprises grapple with social pressure—including pressure from their own workforce—to make responsible choices in an ever more complex world.
“In the past 10 to 20 years, there’s been a lot more emphasis on corporate social responsibility,” he notes. “Customers are willing to pay a premium for a product that adheres to certain ethical norms. And workers would even take a pay cut to work for a firm that ‘does the right thing.’ And firms have started to realize this.”
It remains to be seen if companies will step up to act ethically in providing transparency—and retraining—to workers skittish about AI. But as AI reshapes the business ecosystem, one thing is clear: Right or wrong, it makes good business sense to include workers in the process.
Wind turbines generate electricity without using fossil fuels or producing particulate matter pollution, but they do create waste. Though they can last as long as 25 years, turbine blades cannot be recycled, piling up in landfills at the end of their life. Siemens Gamesa’s RecyclableBlade can be broken down into its raw materials at the end of its life.
AI holds great promise to increase the quality and reduce the cost of health care in developed and developing countries. But one obstacle to using it is patients don’t trust it.
Evidence suggests that most companies are still struggling to build data literacy, even after they’ve identified it as critically important: just a quarter of employees report feeling confident in their data skills. Here are five strategies to help companies improve.