The Impact of AI Integration in Personal Finance Education

Last registered on March 05, 2026
View Trial History

Pre-Trial

Trial Information

General Information

Title
The Impact of AI Integration in Personal Finance Education
RCT ID
AEARCTR-0017918
Initial registration date
February 27, 2026

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
March 05, 2026, 8:45 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Country
United States of America
Region

Primary Investigator

Name
Hakan Ozyilmaz
Affiliation
Stanford University
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Annamaria Lusardi
PI Affiliation
Stanford University
Contact Investigator
PI Name
Daniel Chi
PI Affiliation
University of Nevada Las Vegas
PI Name
Tim Kaiser
PI Affiliation
RPTU University Kaiserslautern-Landau
PI Name
Jodi Letkiewicz
PI Affiliation
CSU Northridge
PI Name
Andrea Sticha
PI Affiliation
Stanford University
PI Name
Caroline Zhnag
PI Affiliation
University of Nevada Las Vegas

Additional Trial Information

Status
On going
Start date
2026-01-20
End date
2027-05-31
Keywords
Education, Finance & Microfinance
Additional Keywords
JEL code(s)
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
This study evaluates whether adding structured artificial intelligence (AI) tools to a college personal finance course improves students’ financial knowledge, decision-making skills, and financial behaviors. We conduct a randomized controlled trial across multiple semesters at two large public universities. Entire course sections are randomly assigned either to follow the standard personal finance curriculum or to use an enhanced curriculum that includes a series of structured, educational AI chatbots designed to guide students through real-world financial scenarios such as budgeting, credit card use, loans, investing, and retirement planning.

The study includes approximately 2,600 undergraduate students across about 70 course sections. Students complete surveys before the course, immediately after the course, and at follow-up intervals to measure changes in financial knowledge, applied financial competence, attitudes, and behaviors. With student consent, survey data are linked to academic records and selected credit-bureau measures to assess longer-term financial outcomes.

The goal of the study is to understand whether integrating structured AI tools into a semester-long college course improves students’ financial decision-making and overall financial well-being, and to inform broader discussions about the role of AI in higher education and financial education policy
External Link(s)

Registration Citation

Citation
Chi, Daniel et al. 2026. "The Impact of AI Integration in Personal Finance Education." AEA RCT Registry. March 05. https://doi.org/10.1257/rct.17918-1.0
Sponsors & Partners
Experimental Details

Interventions

Intervention(s)
The intervention consists of incorporating a series of ten custom-built AI chatbots into the existing course curriculum. Each chatbot is designed to guide students through interactive financial decision-making exercises that mirror real-world situations, such as setting financial goals, budgeting, managing credit cards, choosing loans, understanding mortgages and insurance, investing, assessing risk, and planning for retirement.

The AI tools are embedded within otherwise standard personal finance courses that are delivered either in person or online. In sections assigned to the intervention condition, students complete structured chatbot-based exercises aligned with course topics throughout the semester. These exercises are integrated into the course requirements and contribute to students’ course grades. The chatbots are designed to provide personalized, step-by-step guidance while maintaining pedagogical guardrails to ensure clarity and accuracy.

Course sections assigned to the control condition follow the same core curriculum and cover the same financial topics but do not incorporate the AI chatbot exercises.

The intervention does not replace instructors or lectures; rather, it supplements traditional instruction with structured, interactive AI-based learning activities.
Intervention Start Date
2026-02-20
Intervention End Date
2027-05-31

Primary Outcomes

Primary Outcomes (end points)
The primary outcomes are financial competence and financial knowledge.
Primary Outcomes (explanation)

Financial competence is measured using structured course-based exercises that assess students’ ability to apply financial concepts to realistic decision-making scenarios, including borrowing, saving, investing, insurance, and intertemporal choices. We construct both domain-specific and overall competence indices.

Financial knowledge is measured using an 8-item standardized financial literacy module covering core concepts such as interest compounding, inflation, and risk diversification.

Secondary Outcomes

Secondary Outcomes (end points)
Secondary outcomes include survey-based measures of financial behaviors, financial well-being, AI literacy and reliance, academic performance, and credit-bureau outcomes.
Secondary Outcomes (explanation)
Survey-based financial behavior outcomes include indices capturing budgeting practices, savings, financial inclusion and investment participation, credit exposure, use of alternative financial services, financial fragility, perceived financial capability, financial trust, and sources of financial information (including AI-based sources).

AI-related outcomes include an AI financial literacy index (based on correct responses to AI-related financial questions) and an AI reliance index measuring frequency and reliance on AI tools in financial decision-making.

Administrative academic outcomes include course grades, term and cumulative GPA, enrollment status, and other academic records.

For consenting participants, credit-bureau outcomes include measures of credit utilization, revolving debt dynamics, new credit activity, delinquency, and overall credit exposure.

Experimental Design

Experimental Design
This study uses a cluster-randomized controlled trial design. The unit of randomization is the course section. Across three academic semesters at two large public universities, approximately 70 personal finance course sections are randomly assigned to either a control condition or an AI-enhanced condition.

Sections assigned to the control condition follow the standard personal finance curriculum delivered by the instructor. Sections assigned to the AI-enhanced condition follow the same core curriculum but additionally incorporate ten structured AI-based learning exercises integrated throughout the semester.

Randomization is stratified by university, semester, and delivery mode (online versus in-person). In cases where an instructor teaches multiple sections in the same semester, assignment is balanced within instructor to ensure comparability across sections.

All enrolled students in participating sections are invited to participate in the study. Outcomes are measured using survey data collected before the course, immediately after the course, and at follow-up intervals. Administrative academic data and credit-bureau data are linked for students who provide consent.
Experimental Design Details
Not available
Randomization Method
Randomization is conducted by computer using a pre-specified Stata script prior to the start of each semester.
Randomization Unit
The unit of randomization is the course section (cluster-level randomization). Entire personal finance course sections are randomly assigned to either the AI-enhanced curriculum or the standard curriculum. There is no individual-level randomization within sections.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
70 sections
Sample size: planned number of observations
2500 students
Sample size (or number of clusters) by treatment arms
1250 Treatment and 1250 Control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Accounting for cluster randomization at the course-section level (approximately 69 sections and 2,595 students, with about 38 students per section), assuming a two-sided alpha of 0.05, 80 percent power, and covariates explaining 30 percent of outcome variance, the minimum detectable effect for the main standardized outcomes ranges from 0.14 to 0.22 standard deviations, depending on the intra-class correlation (ICC between 0.02 and 0.10).
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Stanford IRB
IRB Approval Date
2025-07-25
IRB Approval Number
80623
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information