Human Resources, Non-Human Decisions: The Uneven Future of AI at Work

Understanding managers’ perspectives matters because they ultimately decide how AI enters the workplace from the tools that get piloted to the workflows that remain human-led. These decisions don’t just shape organizational strategy but they also ripple outward, influencing how employees experience and accept AI. 

That makes managerial hesitation consequential. If AI acceptance varies across business departments, or if enthusiasm drops when AI acts autonomously, even well-designed AI initiatives may encounter unnecessary friction. 

To unpack where that friction comes from, Miriam Gieselmann and colleagues focus on two factors central to managerial judgment: business area and AI functionality. They ask whether AI is evaluated differently in human resources (HR) than in finance or marketing, and whether acceptance declines as AI shifts from supporting decisions to fully carrying them out. Across four studies conducted in Germany and the UK, the authors show how these factors shape managers’ perceptions of risk, willingness to invest, and willingness to use AI.

A total of 2,025 participants with management experience evaluated vignettes of AI tools being used across business areas. In the first three experiments (Studies 1, 2a, and 2b), participants examined scenarios describing AI use in human resources (HR) versus finance. In the final experiment (Study 3), the comparison shifted to HR versus marketing. The AI tools described varied in functionality, starting low with just monitoring information and ending high with implementing decisions autonomously. After reading each vignette, participants rated their perceived risk, willingness to invest in the tool (studies 1, 2a, and 2b), and willingness to use the tool (Study 3).

Taken together, the findings paint a surprisingly coherent picture about the conditions under which managers are accepting of AI. 

Firstly, is AI acceptance perceived differently depending on the business area? Turns out, it is. Managers consistently show lower acceptance when AI is used in HR tasks, reporting greater perceived risk, lower willingness-to-invest, and lower willingness-to-use. Gieselmann and colleagues theorize that this may be because HR tasks often require empathy, intuition, or contextual judgment that AI tools naturally lack. Plus, HR departments as a whole deal with sensitive personal data, raising privacy concerns. On the other hand, AI was more accepted in business areas that were perceived as more objective and less tied to individual employees, like finance and marketing. Regardless of the root cause, the overall pattern is robust: managers show the most hesitation when AI gets involved in the human side of business. 

Second, does acceptance shift when AI becomes more functional? The findings show that AI autonomy is a clear tipping point. Acceptance declined sharply when the AI tool implemented decisions on its own without human review, regardless of the business area (perceived risk rose and willingness-to-invest dropped).

These sentiments aren’t confined to managers with basic leadership experience. Even senior decision-makers like CEOs and department heads from the companies surveyed in study 3 replicated the core findings: lower acceptance in HR, and lower acceptance for autonomous AI.

With mixed acceptance of AI usage at work, playing it safe by simply avoiding AI tools altogether to prevent organizational friction wouldn’t be ideal. There are growing ways to mitigate friction and increase acceptance. In an era of rapid AI research and development, companies should instead embrace AI tools in a way that works for them. For one, these findings suggest that including human oversight in AI systems would enhance acceptance. On top of that, AI tools should be strategically implemented by business areas—prioritizing departments where tasks are farther removed from employees’ personal stakes can leverage AI’s benefits while keeping the human side of the company human.

These findings open up several important questions about how AI acceptance at work develops and how it might change over time. Future research could clarify what it is about HR tasks that drive low AI acceptance: is it privacy? Fairness? The interpersonal nature of HR decisions? Gieselmann and colleagues also note that real-world experience with AI might shift acceptance patterns over time which paves the path for research that focuses on the longitudinal perspectives of AI use at work. Perhaps as we become more accustomed to AI, we wouldn’t think twice about it.

Managers’ acceptance of AI is not uniform. It varies by business area and drops sharply when AI systems act without human oversight. Because managers ultimately decide which tools are piloted, scaled, or sidelined, these perceptions carry real consequences for how and whether AI ends up taking hold in organizations. The findings suggest that AI’s success at work will hinge less on raw capability and more on careful implementation that preserves human control and respects sensitive business domains like HR. As AI becomes more powerful, understanding these boundaries may be just as important as pushing them.

1. Gieselmann, M., Erdsiek, D., Rost, V., & Sassenberg, K. (2025). Do managers accept artificial intelligence? Insights into the role of business area and AI functionality. Journal of Economic Psychology, 108, 102804. https://doi.org/10.1016/j.joep.2025.102804