Future of Work Lab

🧬🤖 Measuring the Human Experience of AI in the Workplace

📌 What’s This About?

How does it actually feel to work with AI? We study how artificial intelligence is reshaping the employee experience, focusing on job stress, surveillance, collaboration, autonomy, and meaning.

This study was initially inspired by gig work, where algorithmic control, reduced autonomy, and emotional exhaustion are well documented. While the main sample focuses on more traditional sectors, the emotional patterns we measured—especially around stress, surveillance, and purpose—were originally designed to test whether gig-like conditions were spreading elsewhere. In that sense, gig work may be the canary in the coal mine for how AI is reshaping the emotional reality of labor.

Using two short but revealing scales (Task Level Experience Scale and Reflective Work Experience Inventory), the study captures the day-to-day realities and deeper reflections of workers navigating AI's rise.

💡 Why It Matters

Much of the existing research on AI in the workplace emphasizes productivity outcomes. This study centers on people. By surfacing how workers feel about AI, not just what they do, the goal is to support more ethical and human-centered design and policy choices.

🚀 Where It’s Headed

After completing a pilot study with 17 participants to refine the measures, we launched a broader survey across multiple platforms.

In April 2025, we successfully collected responses from over 300 workers across sectors. This expanded dataset is now being analyzed to uncover patterns in AI’s impact on autonomy, stress, collaboration, surveillance, and purpose.

The full dataset, codebook, and working paper are published on the "Measuring the Human Experience of AI in the Workplace" OSF project page.

🤝 Get Involved

• Read the Working Paper: Read it via SSRN
• Browse Data and Code: Downlaod it via OSF
• Curious how this maps to your team? Send an email hello@fowlab.net.

🔬 What We’re Learning: Insights from the Study

Although gig workers weren’t included in the study sample, the emotional and structural patterns commonly associated with gig labor (like surveillance, reduced autonomy, and eroded meaning) appeared consistently across service, civic, and knowledge roles. This suggests that what began in the gig economy is now diffusing into more traditional jobs. These findings validate the study’s original hypothesis and elevate gig work as an early indicator of AI’s broader impact on emotional life at work.

Our research unfolded in two phases: a small-scale pilot study and a large-scale survey. Below are the most compelling patterns we’ve observed so far.

“The real fear isn’t losing jobs to AI. It’s losing agency over how it’s used.”
Preliminary finding, Future of Work Lab

📊 What We Learned from the Pilot Study

• Efficiency Gains: Participants felt more productive and completed tasks faster with AI support.
• Reduced Connection and Trust: Many reported weaker interpersonal connection and less clarity about their roles on the team.
• Surveillance Concerns: Some saw AI as helpful, while others felt increasingly monitored by algorithmic systems.
• Subtle Shifts in Meaning and Autonomy: Work moved faster, but some experienced a dip in purpose or control.

These initial insights formed the foundation for the expanded dataset (n = 300+), which was designed to test and deepen these observations.

📊 What We’re Seeing in the Full Dataset

• AI Boosts Efficiency But Not Always Meaning: Speed gains were common, but some felt work lost purpose.
• Trust and Team Dynamics Are Eroding: Frequent AI use correlated with lower trust and cohesion.
• Surveillance is a Growing Concern: Even assistive tools felt intrusive to some workers.
• Autonomy is Being Reframed: Some found agency in deciding how to use AI, even if not whether to.
• Emotional Ambivalence is the Norm: Workers are navigating excitement, fear, and uncertainty all at once.

These themes will be explored in detail in the forthcoming working paper, to be published alongside the dataset and code on the Open Science Framework.

🔎 Hidden Findings: What’s Often Missed

• The Uneven Reality of AI at Work: Despite headlines declaring AI a universal disruptor, our findings paint a more uneven picture. Several participants, especially those in unionized or labor relations roles, shared that AI has not touched their day-to-day work at all.
• AI Improves Output but Erodes Team Dynamics: While productivity increased for many, AI use often correlated with reduced team cohesion and lower interpersonal trust, especially in collaborative roles.
• The Real Fear Is Losing Agency: Most workers were not worried about being replaced by AI. They were worried about being left out of decisions about how it is used. The anxiety was not about automation, but about losing control over their role and voice at work.

These findings underscore the need for more nuanced, worker-centered approaches to AI design and implementation. Effective strategies should invite participation, preserve autonomy, and build trust.

🧪 Future Plans

• Practitioner Summary: A 5-page white paper drawing from the research, providing practical guidance for leaders, managers, and HR teams, focusing on human-centered organizational change (HCOC) in real-world AI applications.
• Journal Submission: After releasing the preprint, a version will be prepared for peer review in academic journals.
• Optional Phase 3: Qualitative interviews or longitudinal follow-up with participants to explore deeper insights and track changes over time.

⚖️ Research Ethics and Compliance

This study is being conducted independently and adheres to established ethical guidelines for research with human participants. It was originally submitted for IRB review at Rutgers University but has since been withdrawn. The researcher has completed formal ethics training, including:

• Social and Behavioral Research (CITI Program)
• Technology, Ethics and Regulations (CITI Program)
• Research Security and International Collaboration (CITI Program)

All participant data is de-identified and managed using open science and privacy best practices.