At the heart of the AI is a fear that it could become the first technology in human history to lead to fewer jobs for workers and permanently higher unemployment.
The reason this would be so devastating goes beyond simply its financial impact on families. The harm from a dramatic AI-induced increase in mass unemployment would not be cured simply with Universal Basic Income, a policy that shares the gains of AI by simply cutting equal checks to ensure basic sustenance for all. Such an approach fails to address the dignity and fulfillment humans achieve through work and contribution and the need to target such efforts to create a better society.
If an apocalyptic AI job-loss scenario comes true, we do not need to give up on the fundamental goal of building a full-employment economy that centers on work and the purpose it can bring. A far better path is to offset AI-induced job loss with the elevation and expansion of millions of desperately-needed jobs which provide dignified care, preventative health services, educational opportunities, counseling for those with mental health and addiction issues, and navigation services for individuals facing barriers to work due to disability, past incarceration, or long-term unemployment.
In other words, our response to the AI revolution must be to support what I call “double-dignity jobs,” in which human workers receive the dignified compensation they deserve in exchange for caring and providing dignity to other humans.
Double-dignity jobs primarily involve essential human-touch tasks, making them unlikely to be replaced by AI. For instance, research conducted by Anthropic found that health care support and personal care support jobs have some of the lowest risks of automation. MIT’s David Autor has similarly found that these types of care jobs are hard to automate because they involve non-routine, interpersonal, and dexterous tasks. Intuition confirms this analysis. Will most Americans want AI-powered robots to dress, shower, wipe, and comfort a parent with dementia? Or have their kids forgo learning with other children to sit alone with AI tutors? Or instruct teens struggling with bullying and suicidal thoughts to seek counseling only from Claude or ChatGPT?
The problem is that if AI does lead to massive job loss, simply informing job seekers of the resilience of these job categories offers no cure to the sweeping harm of permanently higher unemployment. After all, even with the growing demands of an aging population and increases in intellectual and developmental disabilities (IDDs), larger numbers of these jobs are already poorly paid. Take just one example of many: direct service professionals who help adults with autism. These workers can mean everything to the families they support, yet they are paid so poorly that they face 40% turnover rates. If AI causes higher unemployment, creating a flood of job seekers to chase such jobs will only depress wages, benefits, and opportunities. That is not progress.
This is why the first step in addressing potential AI-induced job loss must be to ensure that such care economy, counseling, and education jobs, which are essential to providing dignity and care to our loved ones, are themselves treated with dignity in terms of the wages, benefits, worker voice, and career mobility they receive.
The second step is to dramatically expand those jobs to meet critical unmet needs. Consider the following: Only a small fraction of the 17 million Americans with IDDs or cerebral palsy receive the individualized support they need. Among the 63 million Americans who provide ongoing care to a family member, a significant share are forced to give up jobs, cut hours, or leave careers because they cannot afford skilled care for loved ones. Eight million students attend schools without a single college or career counselor, while some states have only one school psychologist for every several thousand students. Rural America cannot fill basic medical roles. Or take childcare. There are too few decently paid early-childhood educators to make affordable and universal childcare possible. And we need an army of navigators to help people apply for government benefits, overcome long-term unemployment, and navigate re-entry from incarceration or addiction.
Our societal demand for these kinds of workers is high. And as a society, we must both invest in workers capable of meeting this demand and ensure their services are affordable and available to those in need.
There are several ways we can achieve this in the AI era. For instance, AI tools themselves could be used augment the capacity of these missing jobs to both enhance the economic dignity of millions while increasing returns to society in terms of lower costs of hospitalization, safety net spending, and higher workforce participation.
If AI leads to unprecedented net job loss, one of our strongest paths to maintaining full employment is to create a dedicated tax stream from AI productivity gains to create millions of needed, double-dignity jobs. We will also need New Deal-style jobs in construction and infrastructure to make us more efficient and resilient to natural disasters. But if widespread AI adoption results in a structural net job reduction, a Second New Deal jobs strategy must have this new focus on a permanent expansion of care, counseling, education, and navigator jobs that will last for the long haul.
No one should assume that double-dignity jobs will be the choice of most jobseekers. It should not be seen as a simplistic “coder-to-nurse” strategy. But by adding millions of such jobs, it will benefit all workers. How? A tighter labor market should strengthen wages and worker power across sectors. And unlike policies like Universal Basic Income, this strategy hews more closely to American values. As the first Labor Secretary Frances Perkins once described, when explaining why FDR chose job programs over just sending checks: “Americans want, above everything, to work and contribute.”
There is no question that varying projections of timing and severity of job loss caused by AI make designing such a major job creation strategy an unprecedented challenge. Policymakers today must design plans that both reflect the reality of 4.3% unemployment today and the possibility of major AI-induced joblessness tomorrow. Uncertainty, however, does not have to dictate a wait-and-see approach.
We can start by creating A new fund for double-dignity jobs to meet vital local and regional needs. Such a new job-creation policy could start with moderate funding, but by requiring states and local governments to stand up such programs now, it would lay the groundwork for quick acceleration if AI causes more severe job loss. We can further prioritize those parts of an affordability agenda – like childcare and home care – that also create large numbers of AI-resilient, double-dignity jobs.
AI does not ever have to deny us an economy based on work, contribution, and dignity. Expanding and elevating dignified work to meet our society’s unmet needs provides one of our most promising paths.
The post We Need to Invest in the Dignity of Work for the AI Era appeared first on TIME.
