Sanchaita Hazra
sanchaita.hazra[at]utah.edu
Department of Economics
University of Utah
I am a fourth-year PhD student in Economics at the University of Utah. I am co-advised by Haimanti Bhattacharya and Subhasish Dugar. I also actively collaborate with
Allen Institute for Artificial Intelligence.
My research interests are behavioral economics, experimental economics (Lab/Field), applied microeconomics, and artificial intelligence (AI). My work focuses on applying the methodologies of experimental economics in neoclassical economics to explore and gain deeper insights human decision-making. I also study the human perception and trust in AI-assisted decision-making tasks such as in information gathering, detecting lies, and scientific writing.
My PhD committee includes my advisors, Gabriel A Lozada (UUtah), Daniel Martin (UCSB), and Chris Callison-Burch (UPenn).
In past, I worked as a statistician at DeepFlux (now accquired by Pivot Roots) and a research assistant at Indian Statistical Institute Kolkata.
In 2021, I was also a Lecturer of Economics at Women's Christian College, University of Kolkata. I founded Alankar, a women-run online jewelry brand fostering positive social impact on employbility.
|
CV  | 
LinkedIn
|
Published Works
|
Position: Data-driven Discovery with Large Generative Models
with Bodhisattwa P. Majumder, Harshit Surana, Dhruv Agarwal, Ashish Sabharwal, Peter Clark
Published, July 2024
International Conference on Machine Learning (ICML), 2024
pdf
A practical first step toward an end-to-end automation for scientific discovery. We posit that Large Generative Models (LGMs) present an incredible potential for automating hypothesis discovery, however, LGMs alone are not enough.
|
To Tell The Truth: Language of Deception and Language Models
with Bodhisattwa P. Majumder
Published, June 2024
North American Chapter of the Association for Computational Linguistics (NAACL, Oral), 2024
pdf
We analyze a novel TV game show data where conversations in a high-stake environment between individuals with conflicting objectives result in lies in the presence of an objective truth, a distinguishing feature absent in previous text-based deception datasets. We show that there exists a class of detectors with similar truth detection performance as humans, even when the former accesses only the language cues while the latter detects lies using both language and audio-visual cues. Our model detects novel but accurate language cues in many cases where humans failed to detect deception, opening up the possibility of humans collaborating with algorithms and ameliorating their ability to detect the truth.
|
Experience, Learning and the Detection of Deception
with Priyodarshi Banerjee and Sanmitra Ghosh
Published, July 2023
Journal of Economic Criminology
pdf
Deceptive communication or behavior can inflict loss, making it important to be able to distinguish these from trustworthy ones. This article pursues the hypothesis that repeated exposure or experience can cause learning and hence better detection of deception. We investigate using data culled from events in a TV game show. Decision-makers in the show repeatedly faced situations where they had to correctly identify an individual from within a group all claiming to be that individual. Increased experience reduced average detection error in the sample. Analysis of the data suggested this relationship was significant and driven by learning.
|
Working Papers
|
The Good, the Bad, and the Ugly: The Role of AI Quality Disclosure in Lie Detection
with Bodhisattwa P. Majumder, Haimanti Bhattacharya, and Subhasish Dugar
Job Market Paper
pdf
We investigate how low-quality AI advisors, lacking quality disclosures, can help spread text-based lies while seeming to help people detect lies. Participants in our experiment discern truth from lies by evaluating transcripts from a game show that mimicked deceptive social media exchanges on topics with objective truths. We find that when relying on low-quality advisors without disclosures, participants' truth-detection rates fall below their own abilities, which recovered once the AI's true effectiveness was revealed. Conversely, high-quality advisor enhances truth detection, regardless of disclosure. We discover that participants' expectations about AI capabilities contribute to their undue reliance on opaque, low-quality advisors.
|
Uneven Trust in LLMs: Beliefs About Accuracy Vary Across 11 Countries
with Marta Serra-Garcia
Large Language Models (LLMs) can efficiently provide users with information, but this information may not always be accurate. This paper investigates human understanding of an LLM (GPT-4o) as an information source. In a large-scale experiment with over 2,900 participants across 11 countries, participants were incentivized to evaluate the accuracy of answers to 100 policy-relevant, factual questions, provided by the LLM. Across most countries, individuals show a limited ability to assess LLM accuracy and consistently overestimate this ability. Compared to most countries, participants in the US are more pessimistic about the accuracy of LLMs and less overconfident in their ability to detect LLM inaccuracy, showing significant differences between Americans and other countries in their interaction with globally-available generative AI.
|
Position: AI Safety should prioritize the Future of Work
with Tuhin Chakrabarty and Bodhisattwa P. Majumder
Current efforts in AI safety prioritize filtering harmful content, preventing manipulation of human behavior, and eliminating existential risks in cybersecurity or biosecurity. While pressing, this narrow focus overlooks critical human-centric considerations that shape the long-term trajectory of a society. In this position paper, we identify the risks of overlooking the impact of AI on the future of work and recommend comprehensive transition support towards the evolution of meaningful labor with human agency. Through the lens of economic theories, we highlight the intertemporal impacts of AI on human livelihood and the structural changes in labor markets that exacerbate income inequality. To address this, we strongly recommend a pro-worker framework of global AI governance to enhance shared prosperity and economic justice while reducing technical debt.
|
Funding Fanny - Microfinance and Empowerment of
Women in India
with Sanchita Sen
Bachelors Thesis
pdf
Oral presentation at International Conference on Sustainable Development and Education, 2020
Oral presentation at Research Scholar's Workshop 2020, Visva-Bharati
Women make up a substantial majority of India's poor and they are the cruelest victims of the society. Organizing women through Self Help Groups and equipping them to undertake income-generating activities through the formation of microenterprises have created an economic revolution in the country. The paper focuses on the scope and rationale of microfinance in India and how the Self Help Group-Bank Linkage Programme by NABARD has played its part in empowering rural women financially. We find positive increase in loan disbursements, but sheer increase in loan outstanding over a period of ten years.
|
Working in Progress
|
Groups, Incentives, and Detection of Lies
|
Eliminating a lie - can AI help?
|
Adoption of AI-assistance in Scientific Writing
with Sachin Kumar and Bodhisattwa P. Majumder
|
Awards
- [2024] CSBS Graduate Travel Awards, The University of Utah
- [2024] Haskell Graduate Student Research Award of $1,500, Department of Economics, The University of Utah
- [2023] Research Award of $3,000, Global Change and Sustainability Center and the Wilkes Center for Climate Science & Policy, The University of Utah
- [2023] Graduate Student Travel Assistance Award, The University of Utah
|
Talks
- [2024] The Good, the Bad and the Ugly: Effects of AI Quality Information on Detecting Text-Based Lies at NABETech, Seattle.
- [2024] The Good, the Bad and the Ugly: Effects of AI Quality Information on Detecting Text-Based Lies at ESA, Columbus.
- [2024] The Good, the Bad and the Ugly: Effects of AI Quality Information on Detecting Text-Based Lies at Summer School, Soleto, Italy.
- [2024] To Tell The Truth: Language of Deception and Language Models at NAACL, Mexico. Talk, starts from 1:02:00
- [2024] Humans, Artificial Intelligence, and (Text-based) Misinformation at WEAI, Seattle.
- [2023] Humans, Artificial Intelligence, and (Text-based) Misinformation at ESA, Charlotte.
- [2023] Experience, Learning and the Detection of Deception at WEAI, San Diego.
- [2021] Experience, Learning and the Detection of Deception at Behavioral Econ Workshop, UofU.
- [2020] Funding Fanny--Microfinance and Empowerment of Women in India at Intl Conf on Sustainable Dev & Edu.
|
Teaching
- Instructor, Principles of Macroeconomics, Econ 2010, Summer 2024, UofU
- Instructor, Principles of Microeconomics, Econ 2010, Fall 2024, Spring 2024, Fall 2023, UofU
- Instructor, Intermediate Microeconomics, Econ 4010; Summer 2023, UofU
- Instructor, Intermediate Microeconomics, Econ 6010; Summer 2023, UofU
- Instructor, Q-Pod Tutoring; Spring 2025, Spring 2024, Spring 2023, Fall 2022
- Lecturer, Spring Semester 2021, Women's Christian College, University of Calcutta
|
© Sanchaita Hazra
Thanks to Jon Barron for this nice template
Vibrant Kolkata skyline art is from here
|
|