Boshuo Li, Assistant Professor of Business Administration & Ben L. Fryrear Early Career Faculty Fellow
PhD, University of Miami
MS, Peking University
BA, Tsinghua University
Why did you decide to set out on your research topic?
I have been tracking labor market dynamics in the United States for several years, particularly how AI reshapes high-skilled positions. My coauthors and I began this project in 2022 before ChatGPT was released. At that time, scholars in our field were debating how firms should deploy AI to augment human workers rather than automate their tasks in order to protect employee welfare.
I appreciated the insights from that debate, but I was concerned that the discussion was built on an optimistic assumption about corporate purpose. For example, when workers raise workplace concerns, such as filing discrimination complaints or exposing workplace safety problems to the public, those actions easily draw significant media attention and impose indirect labor costs on firms. Firms then face a strategic decision: they can choose to solve these issues by investing more in their workers, or they can look for other ways to reduce their dependence on the workers who raised these issues.
Today it is widely acknowledged that AI is a viable way for firms to reduce labor-related costs, but when we began this project, few people were willing to say openly that firms might have the inherent motivation to adopt AI to replace workers rather than to support them. Of course, both motivations may exist at the same time, but it is important to understand which motivation actually dominates firms’ AI investment decisions when they face heightened labor-related costs. We therefore assembled large-scale archival data to investigate this question.
What were your key findings?
We find a positive association between media labor issue exposure and firms’ AI investments. At the same time, we observe that such exposure does not prompt firms to invest in IT that is unrelated to AI. These results suggest that firms view AI as a distinct and special type of IT that can counteract media scrutiny of their problematic labor issues. We also examine traditional automation investment, which includes technologies that take over routine and repetitive tasks from human workers in industries such as manufacturing and production. Our analyses reveal that when faced with media exposure of labor issues, firms in high-skilled industries tend to engage in AI investment, whereas those in non-high-skilled industries lean toward traditional automation investment. This divergence highlights the unique automation motive underlying firms’ AI investment toward high-skilled labor.
We also find that AI investment is negatively associated with subsequent media exposure of labor issues for firms previously subject to such scrutiny. Meanwhile, we find that when firms face media labor issue exposure, those that pursue AI investment tend to see a reduction in their future labor forces. This indicates that the observed decrease in media labor issue exposure is likely related to AI “eliminating human labor” rather than “solving labor issues,” providing suggestive evidence that firms’ AI investments are predominantly driven by automation rather than augmentation purposes.
How do the findings of your research contribute to your field?
This is interdisciplinary research at the intersection of information systems and strategic management. In these fields, the conversation around human-AI interactions has largely focused on what AI does to the workforce, like whether it replaces human jobs or creates new jobs. Our study takes the opposite view and asks what labor-related factors drive firms toward AI investment in the first place. This perspective reframes AI investment not only as a proactive strategy firms pursue, but also as a reactive strategy firms feel compelled to undertake under cost reduction pressures.
This study also clearly distinguishes AI investment from non-AI IT investment, which is something the field has called for but only a few studies have examined at the firm level. By comparing AI to traditional automation technologies, we highlight AI’s unique automation mechanism such that it particularly targets high-skilled labor in a way that traditional automation technology does not. That distinction is critical to our understanding of the role of AI within the broader IT landscape.
How do your research findings contribute to society (locally, regionally, nationally, globally)?
Our findings suggest that when faced with pressure to address their labor issues, firms are inclined to replace human workers with AI technologies. Given this tendency, it is critical for workers to continually upskill so as to mitigate the risk of potential job displacement. Employees should focus on enhancing their capability to work alongside AI, collaborate with AI, and adapt to the increasing presence of AI; more broadly, employees should think about how to align their individual value with the organizational value. Furthermore, replacing human workers with AI technologies may contribute to rising unemployment rates, making it imperative for policymakers to devise policies that promote responsible AI development, adoption, and investment, ensuring a balance between technological advancement and social welfare.
Aside from your research, what is the most exciting advancement in your field right now?
The rapid advancement of generative AI has been the most notable development since our paper was accepted. What we have witnessed over the past few years closely reflects what we predicted in the paper. So far, AI’s replacement effects are mainly concentrated among high-skilled workers, while non-high-skilled workers are relatively unaffected. But I’m not sure how long this difference will last. This year we have seen AI getting more advanced not only at cognitive tasks but also at coordinating physical jobs, like AI-powered robots. As these technologies become more mature and cheaper, non-high-skilled workers could possibly face the same AI replacement pressure. I think this is a really critical moment for researchers to look closely at how firms actually use these technologies, and for policymakers to consider how to ensure the gains from AI are shared with workers instead of at their expense.