Comments
THE FEAR OF AI - It is now nearly a century since the 1927 release of Fritz Lang’s Metropolis, depicting a worker revolt against the technology and stratification of a futuristic society. Since then, revolt against technology and automation has been a theme of a line of movies and novels (“Luddite Literature”, as they have been termed).
In the real world, though, the pushback against technology has been more muted. In the post World War II period, attempts through the legal system to halt technology have had limited impacts. Unions and other worker associations have had more success in establishing processes for displacement, job reassignment, and reemployment, but rarely have sought to contest the implementation of technology.
The growth of Artificial Intelligence (AI) already is bringing a different response, a pushback to the use of the technology itself, due to feared job loss. The writer’s strike in Hollywood in 2023 was an early effort to halt the use of AI in ways that might displace jobs; the current effort involving workers at Kaiser Permanente health services is another. At Kaiser Permanente, the union is contesting the introduction of AI as a threat to jobs of its mental health professionals, medical administrative assistants, and behavioral therapists.
Will these new challenges to job loss have any impacts? Should they?
The Post World War II Consensus on Technology and Job Displacement
Over the course of technology advances since the 1950s, a widespread consensus set policy on technology and job placement. This consensus viewed technologies in a range of sectors (manufacturing, financial services, health care) as improving productivity and the living standards of workers across society. The consensus recognized that some workers would be displaced. But it saw the answer in a retraining and reemployment system, that could assist displaced workers to find new jobs.
This consensus was not without legal challenges, including a prominent lawsuit challenging technology in agriculture. In 1979, the California Rural Legal Assistance (CRLA) filed a lawsuit on behalf of farmworkers, to halt research by the University of California related to mechanization in agriculture. CRLA challenged mechanization on the grounds that it would displace farmworker jobs. The case wound through the courts for years. In 1989, the California Court of Appeal held that the University should be given flexibility in technology research, not halted by the judiciary.
Other lawsuits over the next decades involved mainly individual workers contesting their own job loss due to automation: a foreman,whose company automated, challenged his own displacement claiming age discrimination; an administrative worker claimed that while her company may have needed to automate, her job loss was the result of race and sex discrimination; a bookkeeper whose department was automated claimed her job loss represented a disability discrimination. In none of these cases was there a challenge to the right of the company to implement technologies replacing jobs.
In the collective bargaining context, unions have invoked the National Labor Relations Act “duty to bargain” to compel negotiations over the impacts of technology in job reductions. Collective bargaining has addressed processes of layoffs, transfers, retraining and severance. Like the legal challenges, though, collective bargaining largely left unchallenged the right of companies to introduce technology.
CEO: C-suite news, analysis, and advice for top decision makers right to your inbox.
Fear of AI’s Job Displacement and Pushback
AI already is different. Although the number of layoffs due to AI has been modest so far, the fear of layoffs has been extensive, and the union and legislative responses rapid.
The 2023 Writers Guild of America (WGA) strike challenged the use of AI in scriptwriting and show production. The strike lasting 148 days, did bring limits on AI use in scripts, and protections for writers to receive credit and residuals for scripts utilizing AI. In the time since, though, writing jobs in Hollywood have declined in number, due to a mix of economic forces. The Guild has continued to push back against AI use. At present, the Guild’s own staff is on strike, part of its demands relating to protections against AI in their jobs.
At Kaiser Permanente in Northern California, the National Union of Healthcare Workers currently is in contract negotiations, with AI as a primary subject of negotiations. Kaiser has described as AI as freeing employees from routine tasks and paperwork, as well as adding to the quality of diagnoses and treatments. The Union pushback is that AI can lead to faulty diagnoses as well as resulting in job losses.
In the California state legislature, several AI regulation bills are moving forward. SB 951 amends California’s Worker Adjustment and Training Notification process to add a 90-day notice requirement for layoffs due to the use of AI. It also requires employers to prioritize job transfers for AI-impacted workers. SB 947 (“No Robo Bosses” ) sets limits on the use of AI in workplace discipline, and requires human oversight of any disciplinary actions.
Other state legislatures also have responded rapidly with statutes prohibiting the use of AI in displacing workers in certain occupations (mental health professionals, community college faculty), establishing limits of workers to train AI replacements, and requiring an “impact assessment” on job displacement with AI uses.
In the legal context, pushback to AI use has started with the human resources functions. Lawsuits filed in the past two years are challenging the use of AI as discriminatory in sorting job candidates and in directing hiring decisions. With the rapid expansion of state-imposed regulations, though, a wider range of lawsuits is likely to follow.
“AI Derangement Syndrome”
In a recent LinkedIn post entitled “AI Derangement Syndrome” Burning Glass Institute economist Gad Levanon observes the rush of policy analysts, worker associations and employer associations, and law firms to respond to AI. Levanon urges caution. The actual AI impacts on job displacement have been minor; most of the AI use has been augmentation in existing jobs. AI does promise “the most significant transformation of knowledge work in a generation, perhaps a century”, Levanon notes, and cannot be ignored. But the scope and nature of this transformation remain both unclear and to be determined. It’s time to step back and respond to AI’s roles and impacts as they develop—varying among sectors and occupations.
It’s good advice, though not likely to be followed. The past few weeks have brought several high profile layoffs attributed to AI, including the layoff of 4000 workers announced in late February by Block. These layoffs in turn have set off a new round of articles and commentary warning of the scale of job replacement. Last Friday’s latest national jobs report showing a loss of 92,000 payroll jobs, will likely fuel additional calls for government action to curb AI use.
Rage Against the Machine
Throughout the series of technological changes in the post World War II period, what stands out is the remarkable resilience of employers and workers. Employers have pivoted and adapted products and services to utilize technology. Displaced workers, in contrast to their Western European counterparts, have set out on job searches, retrained, been willing to go into different fields.
In the 47 years I’ve been involved in the employment field, I’ve never seen such a brutally competitive job environment: tens if not hundreds of applicants for each opening, several rounds of interviews required, constant frustration, stress, rejection. Yet, for the most part, displaced workers, assisted by the public workforce system, are able to move on, take control of their work futures.
It is this dynamism of employers and workers that state governments, the federal government, and the judiciary should not be in rush to curtail. Employers and worker groups in the various sectors, not government, are best positioned to decide what limits should be imposed on AI. They have the detailed sector knowledge to advise how AI can best be utilized in improving consumer services in, say, health care or financial services, and how to balance service improvements with employment.
Since 2023, economist David Autor and colleagues at MIT have been setting out ideas for a “Pro-Worker AI”: how AI can be used to increase wages of lower paid and least skilled workers. It is one of the lines of thinking to be pursued, though it is not without its challenges in implementation, as AEI’s James Pethokoukis notes this past week.
Finally, policymakers should not be distracted by the theories of universal basic income or guaranteed income that have been making a comeback in the past decade. There is no substitute for the role and power of the job. This is true for the great majority of adults, and it is especially true for workers on the margins: welfare recipients, out-of-school youth, ex-offenders, workers with disabilities and/or severe mental health conditions. It may be that AI will follow the path of previous technological improvements in creating more jobs than it destroys. If not, we need to consider alternatives that enable workers to find roles in the economy and society and structure in their lives—values that jobs bring today.
***
Metropolis plays regularly on the Turner Classic Movies (TCM) channel and continues to be a cult classic. Yet, the director Fritz Lang in his later years came to renounce the movie, for its muddled politics and childish economics.
(Michael Bernick is the former Director of California’s Employment Development Department and previously served eight years on the BART Board. He is currently employment counsel at Duane Morris LLP, a Milken Institute Fellow, and a Fellow at the Burning Glass Institute. A leading voice on workforce issues, Bernick focuses on employment strategies for neurodiverse populations. His latest book is The Autism Full Employment Act. He is a regular contributor to CityWatchLA.com.)
