From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms to Foster Dignified Human-AI Interaction

Paper Title

From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms to Foster Dignified Human-AI Interaction

Publication Info

• Topic area: AI's impact on human expertise and dignity in professional settings.
• Keywords: AI-as-Amplifier Paradox, sociotechnical immunity, Human-AI interaction, skill erosion, knowledge workers, automation bias, dignity preservation, expertise retention, chronic harms, professional identity.

Background and Problem

• Problem / challenge: Current AI systems amplify productivity but erode human expertise and professional identity over time, creating asymptomatic harms that escape traditional metrics.
• Significance: These erosive effects threaten the long-term viability of professional skills, autonomy, and dignity, which are critical for high-stakes domains like healthcare and software engineering.
• Motivation and related work: While prior research has focused on AI's short-term impacts, such as automation bias and deskilling, there is limited understanding of AI's long-term effects on workers' expertise and identity. This paper addresses this gap by studying how AI reshapes human roles and proposing solutions to mitigate hidden harms.

Solution

• Proposed approach: A multi-level framework for Dignified Human-AI Interaction, operationalizing sociotechnical immunity to detect, contain, and recover from AI-induced skill erosion and identity commoditization.
• Novelty:
  1. Conceptualization of the AI-as-Amplifier Paradox, highlighting AI's dual role as enhancer and eroder of expertise.
  2. Introduction of sociotechnical immunity mechanisms to address asymptomatic harms.
  3. Development of a framework co-constructed with workers to preserve dignity, expertise, and agency in AI-mediated workplaces.
• Procedure and key techniques:
  • Conducted a year-long longitudinal study with 42 participants in radiation oncology using mixed qualitative methods (interviews, workshops, think-aloud sessions).
  • Designed and deployed a Social Transparency (ST) layer to surface peer rationales and calibrate AI reliance.
  • Co-constructed the framework through participatory workshops, focusing on worker, technology, and organizational levels.

Results

• Concrete findings:
  • AI integration led to initial efficiency gains (e.g., 5% faster planning cycles) but introduced asymptomatic effects such as vigilance drift and diminished intuition.
  • Chronic harms included demonstrable deskilling, reduced autonomy, and dependency on AI.
  • Identity commoditization emerged as workers feared becoming "AI babysitters" or "rubber stampers," eroding their sense of professional dignity.
• Advantage over baselines:
  • The framework provided actionable mechanisms (e.g., AI-off drills, rationale logs, do-not-automate lists) to counteract skill erosion and preserve human expertise.
  • Social Transparency helped calibrate reliance and foster critical engagement, mitigating overreliance and deskilling.
• Experiments / evaluation:
  • Evaluated the framework in healthcare and software engineering domains through participatory workshops, demonstrating its effectiveness in addressing asymptomatic harms and chronic erosion.
  • Participants reported renewed ownership of tasks and improved resilience through targeted interventions.
• Limitations and future work:
  • Findings are based on a single domain (radiation oncology); future studies should test the framework across diverse domains and populations.
  • Mixed-methods research is needed to quantify degradation rates, recovery timelines, and immunity thresholds.

Summary

This paper identifies the AI-as-Amplifier Paradox, where AI simultaneously enhances productivity and erodes human expertise and dignity over time. Through a year-long study in radiation oncology, it documents asymptomatic effects, chronic harms, and identity commoditization caused by AI integration. The proposed Dignified Human-AI Interaction Framework operationalizes sociotechnical immunity to detect, contain, and recover from these hidden harms. By centering workers' agency and expertise, the framework offers practical tools to preserve dignity and resilience in AI-mediated workplaces, with implications for design, research, and organizational policy.