~1 in 3
Cited industry analyses report that about one in three finance-related AI answers is incorrect or misleading. That is background for the problem, not a finding from this project.
NSF Finders Foundry · Planning phase
Critiquing and Learning to Evaluate AI Financial Reasoning
CLEAR-AI is a classroom research platform for grades 9–12. Students audit simulated AI financial advice using algebra, structured critique, and confidence calibration. AI is the object of scrutiny, not the tutor.
This is educational research. CLEAR-AI does not provide financial advice.
Why this work exists
Financial claims are easy to trust and hard to check.
This is educational research with Township of Union Public Schools. CLEAR-AI does not provide financial advice. It teaches students to test a fluent answer before they trust it.
Students treat the AI output as a claim to audit, not an answer to copy.
47%
Nearly half of consumers have used or are considering generative AI for personal finance. Students still receive little structured practice auditing those claims in school.
Algebra
Algebra is already required. CLEAR-AI gives students a reason to use linear and exponential modeling to check interest, cost, and repayment claims, rather than judging whether an answer merely sounds right.
Method
A critique loop students can actually run in class.
The platform presents controlled, age-appropriate AI outputs that contain deliberate errors. Students work through a five-step critique sequence aligned with the Scaffold–Engage–Apply model.
- 01
Detect
Identify flawed math, hidden assumptions, omissions, and overconfident framing.
- 02
Verify
Recalculate the claim with algebra instead of trusting fluent language.
- 03
Explain
State what the model did, why it is wrong, and whom the error could harm.
- 04
Revise
Rewrite the advice so a classmate could act on a corrected recommendation.
- 05
Calibrate
Compare confidence before and after the work with the errors actually found.
Scaffold
Full, partial, or minimal prompts that help students notice errors at the Remember and Understand levels.
Engage
Error-detection checkboxes and an algebraic workspace for Apply and Analyze work.
Apply
A written critique and pre- and post-task confidence logs for Evaluate and Create work.
Co-design Team
Faculty (Technologists, Researchers)
Dr. Daehan Kwak
Technologist
Kean University · Computer Science
Associate Professor of Computer Science at Kean University and Principal Investigator of CLEAR-AI. He earned his Ph.D. from Rutgers University and an M.S. from KAIST, and coordinates Kean’s graduate computer science programs. His research designs applied AI and intelligent systems for transportation, health, and urban settings, and he is a Faculty Futures Fellow at the New Jersey AI Hub.
Dr. Jennifer J. Chen
Researcher
Kean University · Early Childhood Education
Professor of Early Childhood Education at Kean University and Co-Principal Investigator of CLEAR-AI. She earned her doctorate in Human Development and Psychology from Harvard University. She is recognized among the world’s top 2 percent of scientists in the Stanford/Elsevier rankings and has been honored for teaching, mentorship, research, and scholarship.
Dr. Zimeng Lyu
Technologist
Kean University · Computer Science
Assistant Professor of Computer Science and Technology at Kean University. Her research develops evolutionary neural architecture search for time-series forecasting, producing ultra-lightweight models that can run on edge devices. She has applied this work to industrial operations and financial prediction, and is extending it toward cross-disciplinary edge-AI settings.
Dr. Sai Jambunathan
Researcher
Kean University · College of Education
Dr. Sai Jambunathan is an educator, researcher, and teacher educator with over 30 years of experience in education. Her teaching and research focus on early childhood education, developmentally appropriate practices, financial literacy, artificial intelligence in education, teacher learning, leadership, parenting practices and its impact on children, and gifted education.
K-12 Educators
Dana Bobertz
Supervisor of Humanities, 6–12
Township of Union Public Schools
Dana brings 22 years of classroom experience and 3 years of Instructional Coaching experience to the NSF Grant team. She is also an Adjunct Professor at Kean, teaching English Education and Technology in Education, blending classroom expertise with teacher preparation and instructional innovation.
Marjorie E. Hargrave
Business Educator
Union High School
Marjorie Hargrave is a Business Educator at Union High School. She earned a B.S. in Accounting from Morgan State University and an M.B.A. in Banking and Finance from Dowling College. She works to help students find their strengths, trust their own opinions, and build the confidence to reach their full potential.
Sandy Mangarella
K-12 Educator
Township of Union Public Schools
Sandra Mangarella is a 44-year veteran educator, including 23 years with Township of Union Public Schools. She believes great teaching requires curiosity, courage, and continual growth, embracing innovation while questioning it thoughtfully. Her guiding principle for artificial intelligence is simple: AI should amplify human thinking and learning, never replace it.
Dr. Lisa Kardos
Educator and Chemical Engineer
Township of Union Public Schools
Dr. Lisa Kardos is an educator and chemical engineer who earned her Ph.D. in Chemical Engineering from NJIT. Her background includes university instruction and industrial site technical management. She teaches Honors Chemistry, College Prep Chemistry, and AP Environmental Science for Township of Union Public Schools, specializing in secondary science education, data analysis, and student achievement.
Laurence Petras
Educator
Township of Union Public Schools
Larry is an outside-the-box educator who has spent 40 years inspiring students in the Township of Union. A Notre Dame graduate with degrees in sociology and business and a master's in counseling from Kean, he blends diverse experiences into meaningful learning. He champions the "soft skills" that help every student thrive.
Parents
Dr. George Chang
Parent
Dr. George Chang is a higher education technology educator, researcher, and administrator with over 27 years of experience in the STEM field. With a young child in the Union Township public school system, he brings a different perspective as both a parent and a technology expert.
Dr. Pragya Sharma Ghimire
Parent
Dr. Pragya Sharma Ghimire is an exercise physiology educator and researcher whose work focuses on exercise, nutrition, and bone health. As a parent of a child in the Union Township public school system, she brings an important perspective as both a parent and health sciences expert.
Students
Eric Favaro
Biotechnology major · Junior
Kean University · Union High School graduate
Eric Favaro is a graduate of Union High School and is currently a student at Kean University. He is a biotechnology major in his junior year. In high school he completed both an online and an in-person financial literacy course and believes it is a mandatory course for students.
Eva Philippe
Early Childhood Education major · Sophomore
Kean University · Union High School graduate
Eva Philippe is a graduate of Union High School and is currently a student at Kean University. She is an Early Childhood Education major in her sophomore year. She completed the financial literacy course at Union and is excited to be able to spread the important use of financial literacy.