AI and Human Guidance in Labor Conciliation

How to cite

Seira, Enrique. 2025. "AI and Human Guidance in Labor Conciliation." AEA RCT Registry. AEARCTR-0017348. https://doi.org/10.1257/rct.17348-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Enrique Seira*

* Primary investigator

Registry ID
AEARCTR-0017348
DOI
https://doi.org/10.1257/rct.17348-1.0
First registered
Last updated
Status
In development
Trial dates
2025-11-24 – 2026-02-27
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/17348

Abstract

This randomized controlled trial examines whether providing dismissed workers with clear legal information improves outcomes in Mexico’s mandatory pre-judicial conciliation process. Many workers arrive at conciliation with limited knowledge of their rights and the procedures governing severance, contributing to low settlement rates. In partnership with the Mexico City Conciliation Center, we randomize individuals who come to schedule their conciliation meeting into four groups: (i) a control group receiving no additional information; (ii) a short video explaining the conciliation process and basic legal entitlements; (iii) the video plus legally accurate information delivered by a trained law-student enumerator; and (iv) the video plus legally accurate information delivered through a calibrated large language model (LLM)–based interactive tool.

The study tests whether information provision increases worker knowledge, affects expectations, and improves settlement rates and terms. It also compares the effectiveness of human versus AI-based information delivery. Primary outcomes include settlement incidence, settlement amounts, worker understanding of rights, and subsequent legal actions. Results aim to inform scalable interventions to improve access to justice in labor disputes.

Intervention

This experiment focuses on workers who visit the Conciliation Center to schedule a mandatory conciliation meeting with their employer after being dismissed. Workers will be exposed to one of the following randomly assigned conditions:

a) Control group: No additional information provided.

b) Treatment 1 (Basic info video): participants will receive a video explaining the conciliation stage, and their main entitlements under the law. The video will last 2 minutes.

c) Treatment 2 (Enumerator Info): Participants receive basic video info + information about the conciliation process and labor law from trained law students.

d) Treatment 3 (LLM Info): Participants receive basic video info + information via an AI-powered interactive tool.

Experimental design

This experiment focuses on workers who visit the Conciliation Center to schedule a mandatory conciliation meeting with their employer after being dismissed. Workers will be exposed to one of the following randomly assigned conditions:

a) Control group: No additional information provided.

b) Treatment 1 (Basic info video): participants will receive a video explaining the conciliation stage, and their main entitlements under the law. The video will last 2 minutes.

c) Treatment 2 (Enumerator Info): Participants receive basic video info + information about the conciliation process and labor law from trained law students.

d) Treatment 3 (LLM Info): Participants receive basic video info + information via an AI-powered interactive tool.

Design details

Randomization method
Randomization in the computer
Randomization unit
Groups of 4-5 workers
Clusters
More than 1000 clusters
Observations
More than 4000 workers
Arms
More than 250 clusters per arm
Primary outcomes
1) knowledge index 2) Beliefs 3) How informed and prepared the worker feels, and how prepared and informed the conciliator thinks the worker is. 4) Settlement indicator and settlement amount, by type of case 5) Measures of trust and legitimacy of the justice system

Cited by

No citation data available for this registration yet.

Keywords