5 issues to consider as AI reshapes work

While artificial intelligence is changing industries at an unprecedented pace, it also threatens job security, raises questions about fairness, and could deepen inequality.

At a recent conference convened by the MIT Shaping the Future of Work Initiative, a panel of experts discussed how technology is reshaping the relationship between employers, workers, and labor organizations. 

“How do we make sure workers can embrace the wonder of this technology while we’re also addressing the many big worries?” asked Molly Kinder, a fellow at the Brookings Institution.

Here are some of the speakers’ ideas about how companies can balance innovation and job quality. 

Consider which workers are affected by AI and what safeguards they have

The panel stressed the importance of intentional business choices that support long-term workforce stability.

Frida Polli, a visiting innovation scholar at MIT’s Schwarzman College of Computing, emphasized that businesses have the power to determine how AI is used.

“How do firms decide whether they replace humans or augment them?” she asked. “That really is a leadership decision. I know that’s super fuzzy, but at the end of the day, there is no policy that determines whether you decide to replace or whether you augment human beings.” 

Kinder pointed out that protection from AI’s impacts is not evenly distributed across the workforce. In a case study about Hollywood writers going on strike in 2023 to protect their jobs from AI, Kinder noted that many other workers in the entertainment industry, such as freelance illustrators, voice actors, and independent filmmakers, are not part of a union like the Writers Guild of America and feel especially threatened by AI. 

“The vast majority of workers in this country that are going to be impacted by this technology look more like those creatives than they do the Hollywood writers,” Kinder said. She noted that some of the sectors most exposed to AI impacts, such as finance and insurance, also have low union density, leaving them more vulnerable to disruption and displacement. 

Integrate worker input into AI adoption

Panelists stressed that companies should make AI decisions in coordination with those most affected by them. 

“Workers’ input early is critical. There’s a world of difference between saying, ‘OK, what can the front-line folks tell us about what problems we could solve and how this could improve their job quality?’ versus just a pro forma coming to people and asking for their help in implementing something that is actually scary and intimidating,” said Erin Kelly, a professor of work and organization studies at MIT Sloan.

Amanda Ballantyne, the executive director of the AFL-CIO Technology Institute, said that workers are already experts on how technology affects their roles. 

“They know what makes their jobs better, safer, more efficient [and] what makes their jobs worse, less fair, less safe — and everybody is leaving a lot of money on the table if we’re not engaging them in a very serious way in the R&D process,” Ballantyne said. 

Early engagement with employees can lead to AI solutions that genuinely improve job quality and efficiency — and are more likely to be widely adopted. “We’re not only creating better technology that will be better for society but [will] also have better labor market outcomes,” Ballanytne said.

Increase transparency and fairness in hiring and promotion

Polli highlighted the potential for AI to both detect and correct bias in hiring, noting that this is possible only if businesses commit to transparency. While humans who make hiring decisions often leave no clear record of their fairness or bias, algorithmic systems do; the evidence can be found in the statistical relationships between different demographic groups when AI systems make hiring decisions.  

Polli pointed out that this transparency offers an opportunity: Businesses can assess whether their hiring algorithms disproportionately favor one group over another and, if necessary, adjust them to be more equitable. She cautioned that this potential will only be realized if companies are willing to examine and disclose the results. 

Engage proactively with policymakers to shape policies

Polli said that businesses should actively engage in creating AI regulations rather than waiting for government intervention, arguing that businesses should see regulation not as a threat to innovation but rather as a framework that can foster trust and improve outcomes.

She recalled her own experience lobbying for AI transparency laws, including a fair hiring law in New York City, noting that companies have the opportunity to shape regulations in ways that balance fairness with practical implementation.

“I always feel I have to beat the drum of ‘regulation is not anti-innovation,’” she said, pointing to examples like the Fair Credit Reporting Act, which created clear guidelines for financial data use and ultimately helped the credit industry grow. She suggested that similar regulatory clarity for AI could provide businesses with both ethical safeguards and market stability.

The panelists suggested that partnerships between labor, industry, and policymakers can help ensure that AI is implemented in a way that benefits both workers and employers.

Ballantyne emphasized that many state governments are trying to figure out how to regulate AI and are open to input. This presents an opportunity for labor groups and businesses to step in and help shape policies that are practical and equitable.

Consider AI’s impact on worker well-being

AI-driven job insecurity is more than an economic issue — it can have serious consequences for worker health as well.

“Unemployment is a health hazard,” said Kelly. “We have known that since the 1920s.” Job instability can lead to stress, anxiety, and lower productivity, ultimately harming both workers and businesses.

Kinder pointed out that AI could also reshape work in unexpected but damaging ways. Doctors and therapists, for example, could have AI handle the cases or parts of their jobs that are relatively easy. But “if all the easy cases are done by chatbots, maybe your workload is all the really extreme cases, and that’s going to lead to burnout,” she said. 

Jobs are going to change, and “it’s much more likely we’re going to have equitable outcomes for workers if workers have a say in how that technology rolls out,” Kinder said. 

Watch the panel: Industrial Relations at a Technological Crossroad 

Leading the AI-Driven Organization

In person at MIT Sloan

Register Now