by Henrik Steenkamp, Leuphana University of Lüneburg, Germany.
As a 20-year-old student, I often ask what work will look like in twenty years. With AI, that question feels more uncertain. I don’t think humans will “disappear” from work; a more realistic scenario is that AI becomes deeply embedded in many tasks and changes what jobs look like through augmentation and redesign (Gmyrek et al., 2023).
What worries me is not automation itself—it has shaped work since the industrial revolution—but the possible loss of human interaction at work. Loneliness and social disconnection are already treated as a societal concern, and the workplace is one of the main arenas where adults build daily relationships (Office of the Surgeon General, 2023). If AI takes over more micro-interactions—quick questions, informal feedback, small moments of support—people may interact less with each other, even while “communication” increases.
That matters because belonging is a basic human motive across life domains, including work (Baumeister & Leary, 1995). A lot of what makes teams work is not captured in formal processes, but in repeated small moments that build trust: asking for help, admitting uncertainty, and learning out loud. When those bonding moments disappear, psychological safety can drop, and people may become less willing to speak up, learn openly, or challenge questionable outputs—whether they come from humans or AI systems (Edmondson, 1999). In my view, this is how a workplace can look efficient while becoming socially brittle: it may still deliver results, but it becomes less resilient when mistakes happen or when real conflict needs repair.
This risk grows when evaluation and coordination are shaped by AI. Research on algorithmic control shows that data-driven systems can reshape autonomy and make decisions feel less transparent or harder to contest (Kellogg et al., 2020). When people feel watched, scored, or judged by opaque processes, it can become rational to avoid interpersonal risk. Instead of asking a colleague, they may ask AI. Instead of challenging a decision, they may stay silent. Over time, that can reduce the very peer learning and upward feedback organizations need to improve (Edmondson, 1999).
My entrepreneurial idea is to treat this as a design and capability problem, not an individual coping problem. Instead of one-off training, I would build an ecosystem of practices: clear “no-AI zones” for high-stakes conversations (feedback, conflict, layoffs), manager training, and workplace-level interventions that reduce psychosocial risks and strengthen protective factors (World Health Organization, 2022). I see this as aligned with a broader workplace mental health logic: organizations can’t outsource wellbeing to individual self-regulation when the environment is changing quickly (World Health Organization, 2022).
AI can support inclusion and performance, but the competitive advantage will belong to organizations that use AI without losing the human system that makes work sustainable: trust, speaking up, and belonging (Baumeister & Leary, 1995; Edmondson, 1999).
References
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117(3), 497–529.
Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383.
Gmyrek, P., Berg, J., & Bescond, D. (2023). Generative AI and jobs: A global analysis of potential effects on job quantity and quality(Working Paper No. 96). International Labour Organization.
Kellogg, K. C., Valentine, M. A., & Christin, A. (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals, 14(1), 366–410.
Office of the Surgeon General. (2023). Our epidemic of loneliness and isolation: The U.S. Surgeon General’s advisory on the healing effects of social connection and community. U.S. Department of Health & Human Services.
World Health Organization. (2022). Guidelines on mental health at work
