By Zia Saylor

Discussions of Artificial Intelligence’s (AI) impact in the workplace sometimes center on a doomsday scenario of job automation, where workers are replaced by robots and society sees a surge in poverty and unemployment. Movies, television shows, and literature reveal our fascination with this dehumanization of the workforce and subsequent social breakdown. However, examining existing interactions between AI and social and policy structures suggests that AI can be used in conjunction with good policy and worker power to create a more equitable future with increased opportunities for humans to thrive. To achieve this, however, requires stepping away from the present dearth of worker-friendly policies within the workplace, as without policy changes to undo the power imbalance between workers and their employers, AI will indeed manifest itself as a threat to workers’ livelihoods. 

Worker Power, Income Inequality, and AI Presence

Summer and fall 2023 marked a time of significant union strength, with big wins for the United Auto Workers and Hollywood’s Writers Guild of America and SAG-AFTRA unions. Simultaneously, however, union membership across the nation declined between 2022 and 2023, from 12.2 percent of full time workers to 10.9 percent of full-time workers. Declining union numbers results in part from extensive corporate investment in union busting, and has contributed to rising income inequality nationwide as productivity increases outpace worker pay increases, with profits going to the few corporate leaders. This important context frames the setting in which AI will take place, making legislative efforts all the more important. 

Worker-Friendly AI Policies

Early research on the integration of AI in the workplace indicates that increasing collaboration with unions and input from workers would increase the efficiency of AI workplace usage. This finding is hardly surprising - in a year where a CEO suggested that those impacted by the cost-of-living crisis eat cereal for dinner (ignoring that said crisis was induced by corporate profit-chasing to begin with), it is abundantly clear that corporate leadership can be out of touch with the realities faced by many working Americans, both on the clock and outside of work. Focusing on a just transition to AI, in which workers’ feedback is incorporated and workers themselves are supported in the transitioning labor market, should be a legislative priority to ensure that corporate profit-seeking doesn’t overshadow the need for worker inclusion. Legislation to this effect would help not just workers but also employers, allowing for a more equitable and more efficient AI incorporation. 

Beyond incorporating worker feedback, the U.S. Department of Labor highlights in a recent report the need to train AI itself with ethical worker practices in mind. Given that AI incorporates the biases of those who program and train it, it becomes important to ensure that when developing AI, it is created with a worker-friendly orientation. 

Lastly, the fair distribution of productivity gains also deserves increased regulation in the era of AI. Given that the first study on AI workplace integration found an increase in worker productivity by 14 percent, federal legislation is needed to ensure that such productivity increases are reflected in worker compensation and not simply pocketed by senior leadership. From 1979 to 2022, worker productivity grew 4.4 times more than worker compensation, revealing that corporations will not fairly compensate workers without government regulation, and suggesting that a similar pattern would occur with soaring AI productivity gains. 

(Source: Economic Policy Institute)

Job Market Growth

In adapting to AI in the workplace, new positions will be needed for coding, overseeing, testing, and managing the AI programs. Behind the scenes, human inputs such as text, image, and interaction will be necessary to construct AI models, but regulation should ensure that doing so requires fair compensation of those involved. Even after integration within the workplace, human interactions will continue to shape AI evolution, and as a result, all those who interact with it should be granted a “share” of its value, to be paid out upon exit from a workplace. This model differs from current company-ownership of all workplace goods, but is necessary to facilitate healthy AI growth. Requiring shared AI ownership would also disincentivize overuse of AI as a human replacement, as it would create additional costs for companies. 

Finally, to encourage sustainable and non-exploitative worker treatment, federal and state legislation should ensure adequate compensation for those laid off as a result of AI workforce shifts and provide re-positioning training and support. An example of potential redundancy models comes from the U.K., where redundant workers (those let go through due to no longer needing a role) have rights including the right to being offered a “a suitable alternative role” within the place of employment if one exists and is open, or alternatively, provided with compensation that varies based on age, duration of employment, and average salary while employed. The U.S. currently does not require any such compensation or severance, and as such, only 55 percent of workplaces have written severance policies, even as 88 percent offer some form to those laid off due to redundancy as a form of goodwill. Meanwhile, Americans nationwide experiencing job loss must navigate unemployment insurance (UI), a system that varies by state in amount and duration of benefits receipt, resulting in an unnecessarily-complex web of bureaucratic intricacies. Indiana ranks among the bottom states in terms of amount of compensation that can be claimed while on UI – a mere $390 a week. Fragmented worker safety nets such as incomplete severance and lackluster UI reflect a society that deprioritises worker well-being, something that must change in the future workplace.

Training and the Future of Working Americans

Given the rise in AI-influence technology and AI in workplaces, state and federal digital equity investments should offer individuals the opportunity to engage with and learn about AI in jobs training programs. Pre-existing vocational programs should increase offerings to include supplemental programming on the implementation of AI within even hands-on fields, ensuring that no workers are left behind and that AI does not become a tool of elite cooptation. Within Indiana, the Indiana Community Action Poverty Institute has urged Indiana Broadband to increase the accessibility and equity of AI trainings within their proposed Digital Equity Plan.

The framing of AI as a harm to workers across the country blames AI for anti-worker biases and income inequalities that we as a society have perpetuated through legislative absence. While AI has the potential to exacerbate this reality, with proper regulations and policy shifts, it can also be a force to undo harm inflicted on workers nationwide and create broad prosperity. Whichever road we go down is up to us as workers and empowered citizens.

Want to take action? Digital Equity Investments extend beyond AI.

• Learn more from the National Skills Coalition’s blog post on how to expand access to broadband careers and digital skills including AI here.
• The Institute is also the leader of Indiana’s Skills2Compete coalition which advocates for change around digital equity investments within the state. Join our mailing list here to stay up to date on what additional Indiana-specific action items are coming up.
• Lastly, contact Indiana’s own Senator Todd Young, a member of the Bipartisan Senate AI Working Group, with any thoughts you may have on the future of AI in our workplaces.