When many American workers hear the word artificial intelligence (AI), they associate layoffs, decreased job satisfaction, and tech billionaires getting richer at their expense. they may be right.
But we can also imagine a world where the fruits of AI are instead invested in society. What if every student from K-12 was taught by well-trained teachers in small classes, every patient had the opportunity to interact with a caring nurse, every senior had the opportunity to age with dignity at home or in a quality residential care facility, and everyone could find an affordable therapist when they needed it? And what if care workers were empowered by AI? What if we were all humans, not systems or robots, doing meaningful, well-paying work to rebuild communities and health, supported and rewarded by the outcomes of the AI economy?As technology companies are already targeting cash-strapped care professionals for AI replacement, undermining the humanity and connection we share, here’s a vision of an AI future that puts humans first.
On March 26, the AFL-CIO will host the National Workers First AI Summit to empower workers to help develop policies governing AI in the workplace. The summit comes at a time when discussions about AI and jobs often focus on one question. Will AI destroy jobs, change jobs, create new jobs, or leave jobs largely unchanged? There are doomsayers and apologists on both sides of the debate.
But this framework overlooks a more pressing reality that became clear last month at a regional AI summit hosted by the Cleveland AFL-CIO, Case Western Reserve University, and the Canadian Institute for Advanced Study (CIFAR). Americans are already experiencing a steady decline in the quality of employment. Increased surveillance and surveillance, algorithmic scheduling, and less autonomy, all against the backdrop of stagnant wage growth and a growing affordability crisis. Regardless of whether AI ultimately eliminates millions of jobs, many workers already feel like their jobs are being diminished, or, to use today’s phrase, “coddled.”
AI may accelerate this process, but it is not the root cause. The decline in workers’ bargaining power, the weakening of enforcement agencies such as the National Labor Relations Board (NLRB), and the collapse of union membership all began decades ago, long before modern AI. Meanwhile, institutions such as schools and hospitals that employ and serve millions of Americans remain chronically understaffed. The result is a system with ostensibly strong protection, but with limited real-world impact.
As a society, we have gradually come to accept the simplification of work and the accompanying decline in public and private services. Civil servants are routinely underpaid and overburdened, working in institutions that lack the talent and resources needed to deliver key services. Teachers, nurses, and social workers face increased administrative burdens and ongoing oversight while salaries and benefits have stagnated or declined.
A sector once known for worker autonomy is also starting to feel the pain. Silicon Valley has long been seen as a bastion of professional agency, as tech companies routinely offered generous salaries and perks to software engineers to attract and retain top talent. But those days are coming to an end. Instead, tech companies are using both the achievements and threats of AI to push workers even harder, resulting in longer hours, fewer jobs, and higher expectations for those who remain. Some companies are even adopting China’s infamous “996” schedule, which requires workers to work from 9 a.m. to 9 p.m., six days a week. And engineers are gripped by anxiety about the potential for AI to transform or eliminate their jobs.
they are never alone. We all face a future in which many jobs will fall victim to the AI revolution, and many remaining roles will become unattractive due to the ongoing collapse of regulatory institutions, deteriorating security, and the normalization of management practices that our grandparents would have never dreamed of. It’s no wonder that more than half of Americans surveyed are worried that AI will take their jobs and replace face-to-face relationships. We must not accept these imagined futures.
Human-oriented labor policy recommendations for the AI era
Protecting and enhancing the role of people in nursing care settings
Imagining a better future for workers in the age of AI means protecting the role of people in the workplace. Some jobs should be done by humans. Work that builds human relationships and is important to society, such as teaching and nursing care economics. These professions have long suffered from inefficiencies in their work, faced with overcrowded classrooms, hospital closures, and massive caseloads that are bad for both providers and the people they serve. Instead, people-first policies such as licensing and minimum staffing requirements have the potential to expand the professional workforce and provide rewarding employment while improving the quality of work.
Teach by example. The key to student success is a small teacher-to-student ratio. Small class sizes are a key differentiator for most private schools. However, public schools remain chronically underfunded and have too many students per teacher, leading to poor learning experiences and teacher-student relationships. AI companies didn’t cause this problem, but they could make it worse. A recent report found that students in classrooms using AI feel less connected to their teachers and peers. This is especially concerning given that teacher-student relationships are essential to many student outcomes and many educators have expressed concern about the risks that AI in the classroom poses to student learning.
Rather than replacing teachers with AI, we need to increase the number of teachers. Legislation mandating minimum staffing levels and allocating money for training school staff could put more teachers in classrooms and reduce class sizes. This is not unprecedented. Air traffic control and nuclear power plants require minimum staffing levels, and many states already require maximum numbers of students per teacher in child care and K-12 schools. Similar requirements may be imposed on other care workers. Such workers could still use AI, but in a way that controls and benefits both the workers and the people they serve. In fact, many health systems are already using AI systems to assist with note-taking. This could allow nurses to spend more time with patients.
Develop systems that support training, professionalism and worker rights
Creating and protecting these people-first jobs is not enough. If supply is to match demand, we need training pathways that allow more people to enter these careers at different points in their lifecycles. A hypothetical example of a mid-career transition would be software engineering. Many of the engineers whose jobs are currently threatened by AI have the knowledge and degrees needed to teach math and science, where teacher shortages already exist. Targeted retraining programs and public funding will enable experienced engineers to transition into teaching positions where their expertise can be leveraged.
But training alone is not enough. Engineers choose to enter the technology industry over teaching for a variety of reasons, including differences in income, prestige, and autonomy. Any serious effort to build an engineering-to-education pipeline must address these gaps by improving the incomes and status of people-first professions. It means increasing salaries, ensuring adequate staffing levels, and restoring professional autonomy so that teachers and other care professionals are trusted as professionals. By simultaneously investing in retraining pathways to expand the supply of qualified workers and strengthening these professions to increase demand for their expertise, policymakers can turn the threat of AI exclusion into an opportunity to address long-standing talent shortages in critical public sectors. In these interventions, we find an optimistic vision for the future of AI.
Some of the institutions needed to facilitate such a transition already exist, albeit in reduced form, and could be adapted to serve this new vision. For example, a revitalized NLRB could once again assist workers in negotiating “terms of employment or other mutual aid or protection,” in line with its original mandate. The Fair Labor Standards Act could be amended to include minimum staffing levels for some industries, and the Department of Labor’s Wage and Hour Division could be charged with enforcing them. AI can also be used to support these enforcement efforts and be treated as a way to augment the power of constantly understaffed government agencies.
But other institutions will likely have to be built from scratch. It goes without saying that the United States has never embraced post-employment training and developed a system to link the supply of trained workers with the demand for their skills. On the contrary, training in the United States typically occurs within schools, pre-employment, or on the job. Neither approach will work for mid-career employees in transition.
Fortunately, other countries offer models of lifelong learning to draw on, and there are many local experiments that could be scaled up. New information technologies, if properly managed, could facilitate the adaptation of these models, especially in conjunction with larger worker organizations. After all, it is no coincidence that the most successful models of lifelong learning are found in European countries with strong trade unions. By aggregating the interests of its members and engaging in a productive dialogue with employers, the European Union will become an essential partner in the training process, thereby maintaining the safety of workers and the competitiveness of countries. The revitalization of regulatory and collective bargaining structures may therefore go hand in hand with the development of more robust approaches to lifelong learning.
Creation of a tripartite body to facilitate co-design of AI
Tripartite bodies that bring governments, businesses and trade unions to the bargaining table could support such interventions by identifying additional jobs that require minimal staffing. But it can also serve as a site for productive and participatory design of AI systems themselves. When AI systems are introduced from above, they tend to disadvantage everyone and worsen working conditions. For example, at last month’s Cleveland AFL-CIO conference, utility workers described customer management software that reduces both customer productivity and quality of work life by sending customers down inefficient routes through unsafe streets. Their employer’s general-purpose system simply wasn’t designed for their use case. Employers themselves do not fully understand the details of the downstream use cases required for the design. However, if management and workers can collaborate on the design of bespoke software, they can mutually benefit.
This is not wishful thinking. Computer scientists at Carnegie Mellon University and the UNITE HERE union have co-designed an app to facilitate communication and record-keeping by stewards. This facilitates communication regarding issues such as equipment shortages and minimizes labor-management conflicts. The result is a win-win relationship that opens up a wider range of possibilities, such as leveraging the flow of data from automated equipment to enable trained production workers to solve problems on the shop floor.
High-impact systems used by the federal government, including procured systems, already require consultation with communities affected by the use of AI. However, the post-mortem consultations required by current rules do not give workers the opportunity to shape these systems before they are implemented. Effective participatory design requires understanding and compensating it as work, giving employees a voice, and balancing the tension between co-design in the context of specific use cases and scale of AI systems. Promoting tripartite participation by governments (through grants and demonstration projects in the public sector) can help stakeholders achieve these goals and accelerate the development of worker-centric AI.
Addressing AI and labor migration requires ambitious policies
Americans are concerned about the impact of AI in the workplace. But industry leaders make huge profits while touting the technology’s potential, doing everything in their power to bring innovation to market as quickly as possible, with little regard for those who may be harmed. The gap between elite wealth and average Americans is large and growing. AI has the potential to further widen this divide. Policymakers must meet this moment with an innovative vision for the future of work that puts people first.
