AI’s threat to jobs raises the question of who protects workers

When Congress held a series of hearings on jobs and technological advancement in October 1955, the head of a railroad workers’ organization took the stand to express his fears of automation. “There is anxiety among our employees as they assess the advancement of the new technology,” said WP Kennedy, president of the Brotherhood of Railroad Trainmen. “Will it lead to rising unemployment rather than economic security?”

The same question could have been put to Congress last week at its artificial intelligence hearing. In fact it was.

Sam Altman, CEO of San Francisco start-up OpenAI, testified before members of a Senate subcommittee last Tuesday, urging the government to regulate the burgeoning AI industry. Congressional leaders shared concerns about the threats AI could pose, including the spread of misinformation and privacy violations.

One of their most emphatic concerns has been job displacement: who will take on the responsibility of protecting workers whose jobs could be transformed or even eliminated by generative AI?

Connecticut Democrat Senator Richard Blumenthal stated that his “biggest long-term nightmare” is the job losses AI could cause, before telling Mr. Altman, “Let me ask you what your worst nightmare is.”

“There will be an impact on employment,” replied Mr. Altman. “And I think it takes industry-government partnership, but most of all government action.”

Mr. Altman, like so many other executives unleashing new technologies on the world, has asked the government to assume most of the responsibility in supporting workers through the labor market disruptions caused by AI. It is not yet clear how the government will accomplish this task.

Generative AI, according to a recent estimate of Goldman Sachs. IBM’s CEO already said he expected AI to impact white-collar administrative staff, making up to 30 percent of certain positions obsolete and creating new ones. The White House will host employees on Tuesday for a discussion of their experiences with automation and monitoring technologies in the workplace.

Historically, when automation has led to job losses, the economic impact has usually been offset by the creation of new jobs. Generative artificial intelligence could boost U.S. labor productivity growth by nearly 1.5 percentage points per year over a decade, according to the Goldman report. It could increase annual global gross domestic product by 7 percent. It could give rise to previously unimaginable creative pursuits.

But there will be tremendous instability for displaced workers. Automation has been a major driver of income inequality in America, according to one study from researchers at the Massachusetts Institute of Technology and Boston University. According to their estimates, 50 to 70 percent of changes in the U.S. wage structure since 1980 have been due to loss of income among blue-collar and office workers due to automation.

According to research by Daron Acemoglu, an economist at MIT, parts of the country where robots have been most intensively adopted, particularly manufacturing-heavy parts of the Midwest, are seeing the sharpest declines in employment.

While AI creators tend to focus on the technology’s potential for job creation, many workers will experience painful disruptions as they try to train for and find new positions that pay well and are fulfilling.

“We’ve never been in a period where the scope of automation has such broad potential,” said Harry Holzer, an economist at Georgetown. “Historically, you find something new when your work is automated. With AI, it’s kind of scary that it can just grow and take over more tasks. It’s a moving target.”

Administrative and clerical support workers may be particularly concerned about generative AI, according to the Goldman study. And many of them express all their concerns.

“It’s definitely scary,” said Justin Felt, 41, a customer service representative in Pittsburgh who has worked for Verizon Fios for nearly 12 years. He feels employers haven’t been completely upfront with their employees about the ways they’re integrating generative AI into customer support functions, he said. “It certainly costs us work.”

These technologies are rapidly entering workplaces. BuzzFeed just launched a chatbot that makes recipe recommendations, McKinsey helps customers use AI to fix tech bugs, and accounting firm KPMG uses ChatGPT to generate code. Some economists have therefore started putting forward proposals to protect the workers most likely to be affected.

For example, employees could benefit from a paid time off policy that allows them to take time off from their job to develop new skills. Germany already has a similar program, under which workers in most German states can spend at least five paid days a year on educational courses, an initiative the Labor Minister recently said he plans to expand.

Another possibility is a relocation tax, levied on employers when an employee’s work has been automated but the person has not been retrained, which could make companies more inclined to retrain workers. The government could also provide financial incentives for AI companies to create products designed to improve what employees do, rather than replace them – for example, AI that researches TV writers but doesn’t draft scripts, which is likely to be of low quality. are.

“If the government sets the agenda in developing technologies that are more complementary to humans, that would be very important,” said Mr. Acemoglu. “The industry looks to the government for leadership.”

The government’s previous efforts to support workers during periods of job displacement have had mixed results. a study of Trade Adjustment Assistance, a U.S. government program that provides financial assistance and training to workers who lose their jobs due to trade, found that manufacturing workers who temporarily left the workforce in the early 2000s to participate in the program still were not caught several years later on income compared to workers who lost their jobs but were not eligible for TAA support.

Many economists say employers can also play a role in helping displaced workers.

“Business is always looking to government to deal with job losses,” said Simon Johnson, an MIT professor and co-author with Mr. Acemoglu from the book “Power and Progress”. “But Microsoft and Alphabet — they’re in the driver’s seat, as to where they choose to put their technology assets.”

For example, employees could benefit from apprenticeships and retraining programs with employers. The accounting giant PwC recently announced a $1 billion investment in generative AI, which will include efforts to train its 65,000 employees on how to use AI. ongoing discussion about generative AI

“Some of us walking out of that room knew something had changed,” recalled Joe Atkinson, the company’s chief products and technology officer.

According to Mr. Atkinson expressed fears of displacement, especially as their company explores role automation with generative AI. not be eliminated.

Some tech companies offer employees courses in cloud computing, cybersecurity, and generative AI. Among them is IBM, which also has an apprenticeship program that trains employees, including those without a four-year degree, for high-paying positions in areas such as software development and data science. The C3 AI company offers its 1,000 employees bonuses of $250 to $1,500 for certification in technology subjects, including AI and cloud computing. KPMG is working to train all its employees in the use of generative AI

Community colleges are also intensifying their focus on AI. Miami Dade College has received more than $15 million in grants for its technology programs, with some of the money used to open two centers aimed at preparing students for careers in AI Houston Community College recently awarded a bachelor’s degree in AI and robotics announced, and Southwest Tennessee Community College is working to create an associate degree. The American Association of Community Colleges has launched an AI incubator network aimed at helping faculty teach AI and colleges create AI degrees.

“As Wayne Gretzky once said when asked about his success, ‘I skate where the puck goes,'” said Dennis Natali, a professor at Pikes Peak State College in Colorado who released a plan this year to earn AI certifications. to roll out. “Our college is constantly assessing the labor landscape and preparing to support displaced workers.”

As colleges and businesses scramble to retrain workers, some experts are optimistic about this technology transition. They note that people throughout history have feared technological advancements, but have often taken advantage of them, going back to the Luddites, weavers who protested the mechanization of the textile industry.

But that does not mean that the transition period will be smooth. Michael Chui, an AI expert at McKinsey, pointed out that even the Luddites saw their income stagnate for decades.

“Anyone who loses their job involuntarily – it’s a tough time,” he said. “In some ways the Luddites weren’t wrong about the risk.”