FideAI

FID-086 · Open question

Workplace AI, Durable Skills, and Worker Agency

How does AI-mediated work change workers' skills, decision authority, and ability to adapt, beyond short-term productivity gains?

Why the question remains open

Task throughput can rise while useful judgment, bargaining power, or opportunities to learn decline. A trustworthy-AI research agenda should distinguish benefits to an organization from benefits and burdens experienced by workers and their households.

Working hypothesis

A proposition to test, not a finding.

Worker-controlled assistance with protected learning opportunities will better preserve independent performance than automation optimized only for throughput. Effects may depend on role, prior skill, and how gains are distributed.

Proposed method

How the question could be tested

  • 01Start with consenting adult participants in simulated work tasks; compare unaided, automation-led, and worker-controlled assistance with matched objectives.
  • 02Measure output quality and time alongside delayed unaided performance, ability to detect errors, decision discretion, learning opportunities, and perceived pressure.
  • 03Scope a separate longitudinal organizational study with worker participation to examine role transitions and distribution of benefits; avoid inferring livelihood effects from task experiments alone.

Needed controls

What must constrain the study

  • 01No covert employee surveillance or individual performance reporting to employers. Establish consent, confidentiality, and withdrawal protections.
  • 02Account for prior expertise, job security, accessibility, and selection effects. Pre-register subgroup analyses and follow-up periods.
  • 03Disclose the definition of useful work and whose interests inform success measures. Separate perceived autonomy from actual decision rights.

Relationship to existing work

Preserves FID-046 through FID-050 as faith-context work and extends the broader questions of useful contribution, durable skills, and worker agency.

Expected outputs

Artifacts the work should produce

  • 01A worker-agency and durable-skills assessment protocol.
  • 02A research plan for longitudinal effects and a report of simulation limits.

Open questions

Uncertainties the protocol must resolve

  • 01Which changes in skill and discretion persist after assistance is removed?
  • 02How can workers help govern the study and benefit from its outputs?

Open question

Open work

Primary need: labor research, organizational studies, worker participation, longitudinal evaluation

  • Workers, labor researchers, and organizational scholars to co-design outcomes.
  • Employers and public-interest partners willing to support independent, consent-based research.