AI Education Innovation Fund

AI Education Innovation Fund is sponsored by Wharton AI & Analytics Initiative, University of Pennsylvania. This fund supports the development and enhancement of AI-focused teaching for degree and non-degree learners.

Funded Research - Wharton AI & Analytics Initiative Researchers » Funded Research Since its launch, the Wharton AI & Analytics Initiative has invested in faculty research and educational innovation across the School. These metrics highlight the breadth of our funding programs and the impact they have had in supporting interdisciplinary research, advancing AI and analytics, and fostering new ideas.

Supports research exploring the impact and application of AI across business, industry, and society. Funding may be used for research projects or general research support, including data acquisition, computing resources, research assistance, and, in select cases, matching support for postdoctoral and doctoral researchers.

Applications Accepted: June & December Data Science and Analytics Fund Supports research advancing innovative applications of data science and analytics to address challenges across business and society. Funding may be used for research projects or general research support, including data acquisition, computing resources, research assistance, and, in select cases, matching support for postdoctoral and doctoral researchers.

Applications Accepted: June & December AI Education Innovation Fund Supports the development and enhancement of AI-focused teaching for degree and non-degree learners. Funding may be used to develop new AI courses, enhance AI learning tools, update and expand existing course materials, and support other innovative AI education initiatives.

Talking with Your Hands: How Hand Gestures Influence Communication Authors: Jonah Berger, Giovanni Luca Cascio Rizzo, Mi Zhou Human-Algorithm Collaboration in Gig Work: The Role of Experience, Skill Level, and Task Complexity Authors: Benjamin Knight, Dmitry Mitrofanov, Serguei Netessine Generative AI Shifts Technical Knowledge Production Toward Recombinant Novelty Authors: Simin Li, Neha Sharma The Artificial Intelligence Disclosure Penalty: Humans Persistently Devalue AI-Generated Creative Writing Authors: Justin M.

Berg, Manav Raj, Rob Seamans Experimental Evidence of the Effects of LLMs Versus Web Search on Depth of Learning Authors: Shiri Melumad, Jin Ho Yun Improving Access to Essential Medicines Via Decision-Aware Machine Learning Authors: Jatu Abdulai, Hamsa Bastani, Osbert Bastani, Patrick Bayoh, Angel (Tsai-Hsuan) Chung, Lawrence Sandi, Francis Smart Funded Projects Summer 2026 AI Safety Via Adaptive Defender Agents Enric Boix, Assistant Professor, Statistics and Data Science We propose a framework for AI oversight that casts it as an adversarial game between a worker agent and an overseer agent.

Prior work has shown that RL-trained agents can learn to evade static monitors, raising concerns about the viability of monitoring-based oversight. We argue, however, that when the monitor is adapted in tandem with the worker, it can keep pace with—and counter—evolving evasion strategies. This motivates a shift toward adaptive monitors.

We support this argument with experiments in a simple cybersecurity setting, demonstrating that adaptive defenders can overcome adaptive attackers. Does AI Have a Taste for Discrimination?

Evidence from Algorithmic Hiring Kai Cooper, PhD Student, Operations, Information and Decisions Dean Knox, Assistant Professor, Operations, Information and Decisions This project develops an experimental audit for detecting whether AI hiring systems exhibit taste-based discrimination in recommendations.

In economics, two mechanisms are commonly used to explain group-based disparities in decisions: taste-based and statistical discrimination. In the context of hiring, taste-based discrimination refers to differential treatment arising from a direct preference for or against hiring members of a group.

Statistical discrimination refers to differential treatment arising because group membership is used in combination with signals generated by the decision-making environment to form beliefs about factors relevant for hiring, i.e. an individual’s expected productivity.

Generally, these mechanisms are difficult to distinguish because beliefs are unmeasured by the analyst, however, we may access them directly within modern AI tools, such as LLMs.

Based on a novel conceptualization of these mechanisms as mediated effects in a causal model, I derive a new test of taste-based discrimination that is robust to biased beliefs: among cases where race leaves predictions of productivity or fit unchanged, a found difference in the recommendation reveals taste-based discrimination.

I will construct synthetic resumes with fixed productive characteristics and randomly assigned race-coded names.

Leading AI models from OpenAI, Anthropic, and Google will evaluate each paired resume under two pipelines: (i) a direct AI recruiter that produces both a fit score and interview recommendation, and (ii) a pipeline in which an embedding-based resume-job matching system first produces a fit score that a downstream AI uses, along with the resume, to make an interview recommendation.

The key test focuses on resumes for which the model’s predicted productivity or match score is unchanged across racial cues, where, under the hypothesis of no taste-based discrimination the recommendations produced should also be unchanged in this subgroup.

The project contributes to AI governance, algorithmic fairness, and causal inference by providing a portable design for examining the presence of discrimination in high-stakes AI decision systems.

The Early Diffusion of Embodied AI: Robot Capabilities, Adoption, and Work Exposure Shuyi Dong, PhD Student, Operations, Information and Decisions Lynn Wu, Associate Professor, Operations, Information and Decisions; Director of Embodied AI and Robotics at the Mack Institute Embodied AI integrates artificial intelligence into physical systems that perceive, move, manipulate objects, and interact with people in the physical world.

Despite rapid advances in robotics hardware, foundation models, teleoperation, and simulation, we still have limited systematic evidence on how embodied AI is entering organizational use. This project builds a data-driven map of the early embodied-AI landscape by linking three layers: what embodied AI can do, where they are adopted, and which human tasks are most exposed.

Using publicly observable contracts, deployment records, product releases, firm information, and labor-market data, we will measure the evolution of robot capabilities, identify early adopters across industries and regions, and examine how robotic task capabilities correspond to the structure of human work.

The project will provide an empirical foundation for understanding whether embodied AI is moving toward general-purpose technology or diffusing through specialized use cases. Demand Concentration and Discoverability in AI Search Kartik Hosanagar, John C. Hower Professor of Technology & Digital Business; Professor of Marketing, Operations, Information and Decisions AI assistants are changing how people search for products and services.

Instead of returning a long list of results, they hand the user a short list of recommendations, which means they now decide which options people see. This project asks whether that change concentrates demand on big incumbents and large aggregators at the expense of small and local providers.

By measuring which brands AI assistants surface, where, and for how long, we aim to establish the first empirical benchmark of concentration in AI search and to identify what determines who makes the shortlist. African Digital Infrastructure Governance Atlas Julian Jonker, Assistant Professor, Legal Studies and Business Ethics Africa is in the early years of a digital infrastructure buildout.

Data centers, fiber networks, subsea cable landings, indigenous AI models, and compute capacity are being financed by a mix of public and private sources. These investments promise economic development through connectivity and AI readiness, but their effects depend upon the governance arrangements by which they are financed, regulated, and integrated into local economies.

These arrangements are largely unmapped, making it difficult for researchers to draw lessons about how they influence outcomes. ADIGA is a research program aimed at creating a project-level dataset that captures projects such as data centers, cable and network improvements, and indigenous AI models along with data about how they are financed and which legal instruments govern them.

This information is dispersed across legislative gazettes, regulators and data protection authorities, regulatory filings, industry associations, operator disclosures, news reporting, and informal local knowledge. It ranges from public to inferred to unavailable.

This pilot aims to show the feasibility of creating a full dataset by (a) auditing data availability for a limited set of countries; (b) forging local partnerships that will extend our ability to validate local sources and obtain non-public data; and (c) creating a pilot coded dataset for a subset of projects.

Corporate Footprints: Measuring Firms’ Environmental Impacts at Scale Rongchen Li, Assistant Professor, Accounting Firms are major contributors to climate change and environmental degradation, yet the contribution of individual firms remains largely unmeasured: of the millions of firms worldwide, only a few thousand report their carbon emissions, reporting mandates cover select firms in a few jurisdictions, and self-reported data are aggregate and of questionable credibility.

We develop independent, comprehensive, and localized measures of corporate environmental footprints—covering greenhouse gas (GHG) emissions, air quality, water quality, and biodiversity by connecting high-resolution spatial environmental data with the geocoded locations and financials of hundreds of millions of firms worldwide in a flexible hierarchical Bayesian framework.

The resulting firm-level measures present a comprehensive, independent picture of corporate environmental impacts at global scale—covering the vast majority of firms that existing measures miss—and speak directly to the design of effective climate policy.

Learning to Learn: Algorithmic Design for Effective Education Xufei Liu, PhD Student, Operations, Information and Decisions Gad Allon, Professor, Operations, Information and Decisions Long-term memory is key to deeper learning, yet students differ drastically in how they acquire and retain knowledge.

Current education platforms offer individualized practice at scale to address student heterogeneity, but most approaches optimize engagement, with long-term learning as an afterthought. To address this, we partner with a Chinese language-learning flashcard platform with thousands of active students and ~6 million card reviews per month, where each word is practiced across four skills (reading, writing, definition, and tone).

Using large-scale behavioral data, we (i) structurally estimate individual learning/forgetting dynamics (including differences across skills and content difficulty) and (ii) design an algorithm that personalizes review timing and challenge to improve retention and learning efficiency, while managing the risk of student attrition as sessions become more difficult.

We capture learning across related contextual skills, making it possible to measure both skill-specific learning and skill transference learning. Through this, we suggest review sessions which target words that have the largest leverage in both types of learning, leading to greater improvement of student skill without increasing difficulty and decreasing engagement.

Client-interface Human Capital and Occupational Exposure to Generative AI Xinyu Ma, PhD Student, Operations, Information and Decision Lynn Wu, Associate Professor, Operations, Information and Decisions This project examines whether generative AI exposure affects knowledge workers’ career trajectories differently depending on how their expertise is delivered.

We hypothesize that AI’s impact depends not only on what expertise workers process, but also on how their expertise reaches the people who use it. While some expertise is delivered through codified outputs, other expertise is through direct interaction with clients and partners.

We conceptualize the latter mode of delivery as client-interface human capital: the skills that operate at the interface between expert knowledge and the clients who use it. We examine whether workers whose jobs involve client-interface human capital are less likely to be displaced than otherwise similar workers whose expertise is more easily codified.

Our project informs discussions about the future of AI-mediated knowledge work by highlighting the growing importance of expertise delivered through human relationships.

Tell Me the Truth: Conjoint Analysis in the LLMs Era Ruben Ramirez Salas, Phd Candidate, Operations, Information and Decisions Kartik Hosanagar, Professor of Marketing, Operations, Information and Decisions This project studies whether large language models truthfully explain the product attributes driving their recommendations.

We compare LLMs’ stated rationales with model-internal attribution measures from the same product choices, then test whether more faithful rationales affect consumer persuasion, choice quality, and trust. Who Pays in the New Equilibrium?

Airbnb’s Fee Redesign, Host Exit, and Market Concentration Neha Sharma, Assistant Professor, Operations, Information and Decisions Using a natural experiment in Airbnb and data of all listings in the US across multiple platforms, we study the new market equilibrium: how much of the fee change is absorbed in host payouts vs guests, the differential impact on professional vs casual hosts, and geographies.

School-Year Start Timing and Firearm Deaths Among Children in the United States Dylan Small, Professor, Statistics and Data Science Guns are the leading cause of death in the U.S. among children 1 to 17. Recreations and parks programs can potentially reduce child gun deaths by offering young people more activities and safe places to hang out during the afternoon and evening.

In many places, there are summer recreation and parks program that end toward the end of the summer. In Baltimore, funding for these programs was extended to the first week of school. What effect might such a policy have on a nationwide basis?

To address this question, we will use state of the art causal inference methods to look at the effect of school start time on child gun deaths and what effect the ending date of summer recreation and parks programs has on gun deaths.

Generative AI, Knowledge Organization and Firm Boundaries Oliver Sun, PhD Student, Operations, Information and Decisions Lynn Wu, Associate Professor, Operations, Information and Decisions As generative AI transforms how knowledge is created, shared, and applied, understanding how firms organize specialized expertise has become increasingly important.

This project examines how knowledge organization shapes firm boundaries and how these relationships change in the AI era.

Using a novel project-level organizational dataset, the study investigates: (1) how the distribution of technical and creative expertise influences firms’ production and control boundaries, and (2) how generative AI reshapes these relationships by changing the creation, transfer, and coordination of specialized knowledge.

The project will provide new insights into how firms organize knowledge-intensive work and how generative AI is transforming collaboration, organizational design, and the balance between internal capabilities and external partnerships.

The Implications of Integration of LLMs in Search Advertising Pinar Yildirim, Associate Professor of Marketing This project studies how the shift from keyword search to large language model (LLM) based, conversational search is reshaping the economics of digital advertising.

As users move from terse keywords (“car insurance”) to detailed prompts (an “agreed value policy” for a “classic car used at track days”), platforms observe far richer signals of intent. Whether this enrichment raises or lowers platform advertising revenue is contested: industry analysts warn that answer-first interfaces will erode search advertising, while platforms are racing to monetize conversational data.

I propose to resolve this question empirically by measuring how the concentration, uniqueness, and linguistic complexity of advertiser keyword targeting have changed around the diffusion of LLM search, and how paid and organic search traffic have been reallocated. The analysis pairs a formal auction-theoretic framework with two complementary commercial datasets.

Financing the AI Landscape: A Tale of Tangibles and Intangibles Yao Zeng, Assistant Professor of Finance The build-out of artificial intelligence requires large-scale up-front tangible investments in data centers, power connections, cooling systems, and related physical infrastructure.

Yet the economic returns to these investments are realized primarily through follow-on intangible developments—the training of AI models and the provision of data and compute services—which feature strong network effects and spillovers. We propose and test a framework for financing the AI landscape that highlights this intertemporal interaction between tangible and intangible capital.

Recent industry estimates project total AI-related infrastructure investment on the order of trillions of dollars over the next decade, with more than half of this capital supplied externally through private credit, structured finance, and project-level debt vehicles.

We document a set of stylized facts about the financing of AI infrastructure, showing that debt capacity is critically supported by long-term pre-leasing agreements and concentrated tenant relationships with highly rated hyperscalers.

We then build a model in which short-term debt and pre-leasing contracts transform future intangible growth and network effects into pledgeable cash flows, expanding today’s borrowing capacity for tangible investment.

The model characterizes the optimal capital structure and pecking order of financing the AI landscape, and shows that this financing design can also amplify over-investment and financial fragility when network benefits fail to materialize.

The project advances our understanding of how tangible and intangible capital interact in the financing of general-purpose technologies, offering new perspectives on debt capacity, asset valuation, and the potential for asset-price bubbles in AI infrastructure.

Tissue World Models for Precision Diagnosis and Personalized Therapeutic Prediction Nancy Zhang, Ge Li and Ning Zhao Professor; Professor of Statistics and Data Science; Vice Dean of Wharton Doctoral Programs This project will develop a Tissue World Model: a generative AI framework for building patient- and disease-specific tissue digital twins from single-cell and spatial genomic data.

The model will learn how molecular programs, cell states, spatial neighborhoods, and intercellular interactions shape disease progression and treatment response. Once trained, it will enable virtual genomic perturbations, allowing researchers to simulate how a tissue may respond to changes in genes, pathways, cell states, or microenvironmental signals.

The project addresses a central challenge in precision medicine: translating complex tissue genomics into actionable predictions for diagnosis, prognosis, and therapeutic selection.

By integrating single-cell transcriptomics, spatial transcriptomics, lineage-aware modeling, and virtual perturbation methods, this work will create AI tools for precision diagnostics, personalized medicine, and therapeutic target prioritization in aging, fibrosis, cancer, and other diseases where tissue organization is central to clinical outcome.

Funded Projects Spring 2026 4th Annual Penn-Georgetown Digital Ethics Workshop Brian Berkey, Associate Professor, Legal Studies and Business Ethics Facing Default? AI-Extracted Facial Features and Credit Outcomes Marius Guenzel, Assistant Professor, Finance This project studies whether AI-extracted facial features contain economically meaningful “soft” information that predicts credit outcomes beyond traditional credit bureau measures.

Using a novel dataset linking LinkedIn profile images, labor market histories, demographics, and credit bureau records for nearly 1. 5 million individuals, we apply state-of-the-art computer vision models to generate facial embeddings and evaluate their out-of-sample predictive power for delinquency.

Facial embeddings significantly predict default risk and add incremental value beyond credit scores, income, and education, with particularly strong gains for thin-file borrowers. Planned extensions expand credit outcomes and develop survey-validated personality measures to shed light on underlying mechanisms.

The project informs both the economic role of soft information in credit markets and policy debates surrounding AI-based screening tools.

Workshop on Responsible AI and Data Science for Insurance Pricing Giles Hooker, Professor, Statistics and Data Science The rapid adoption of AI and data science in insurance underwriting and pricing has created a pressing challenge: how to achieve highly accurate risk prediction while ensuring fairness, transparency, and public trust —particularly as climate-related disasters intensify risk heterogeneity across regions and populations.

As insurers increasingly rely on complex models and sensitive data, concerns around bias, accountability, and regulatory compliance have become central to both industry practice and policy debates.

This workshop, co-organized by Professor Giles Hooker (Wharton) and Associate Professor Fei Huang (UNSW), will convene leading researchers, industry practitioners, and regulators to examine frontier methods and governance frameworks for responsible insurance pricing.

The event will explore how predictive accuracy, fairness, and market efficiency can be balanced in AI-driven pricing models, including in high-risk and climate-affected markets, and will foster sustained industry–academia collaboration to support the responsible use of data science in insurance.

Kartik Hosanagar, Professor, Operations, Information, and Decisions As LLMs increasingly mediate consumer decisions, organizations need methods to optimize content visibility while maintaining authenticity. This project develops agentic optimization techniques that identify high-impact content modifications through gradient-guided search across open-weight models.

While optimization mechanisms are becoming well understood, actual consumer interactions with AI-optimized content remain largely unexplored. Perceptions of Fairness in Algorithmic Decision-Making Bethany Hsiao, PhD Student Duncan Watts, Professor, Operations, Information, and Decisions In algorithmic decision-making, it is often mathematically impossible to satisfy competing definitions of fairness simultaneously.

This impossibility makes assessing stakeholders’ fairness preferences a challenging but crucial exercise. To address this, the researchers developed an interactive tool to elicit fairness preferences by projecting the complex decision space into a simple one-dimensional choice: setting risk thresholds.

This design allows participants to directly manipulate algorithm parameters and visualize the resulting tradeoffs between competing fairness metrics. Participants’ decisions enable the researchers to study (1) what Pareto-optimal settings are perceived as fair and (2) whether the design of the elicitation process itself can fundamentally shape how fairness is perceived.

Generative AI and Startup Scaling: Evidence on Reducing Supply- and Demand-Side Constraints J. Daniel Kim, Assistant Professor, Management Startups play a central role in growth and job creation, yet scaling remains a fundamental challenge. On the supply side, young firms struggle to attract and organize high-quality talent, while on the demand side they face severe uncertainty in discovering and sustaining customers.

This project examines whether generative AI reshapes these constraints to scaling. We argue that generative AI enables startups to scale differently by reducing reliance on headcount growth and improving customer discovery and value delivery.

Using a novel panel dataset of venture-backed software startups, we integrate data on generative AI adoption, employee-level hiring and workforce composition, venture capital financing, and real-time website traffic as a proxy for customer demand. Exploiting variation before and after the release of ChatGPT-4, we estimate the effects of generative AI adoption on startups’ employment structure, market traction, and overall productivity.

Whom, When and What to Nudge: Dynamic Personalization via High-Dimensional State-Switching Bandits Eric Bradlow, Vice Dean, Wharton AI & Analytics Initiative Peter Fader, Professor, Marketing E-commerce platforms increasingly use real-time nudges like discounts, urgency cues, and personalized offers to drive conversions.

Yet optimizing these nudges is challenging: user context is high dimensional, and consumers move through unobserved behavioral states like casual browsing, product evaluation, and purchase intent, each responding differently to interventions. We develop a state-switching bilinear bandit model that learns whom to target, when, and with what incentive.

The model combines a Hidden Markov Model to infer latent shopping states, a bilinear structure to capture interactions between user features and nudge content, a spike-and-slab prior for automatic variable selection and a complexity prior to manage the overall size of the parameter space.

We embed this within a forward-looking Thompson Sampling policy that accounts for discounted memory — ensuring that promotions don’t backfire by raising future expectations. Empirically, we validate the approach on synthetic and real datasets.

Our approach offers a scalable solution for dynamic promotion design, outperforms benchmarks that ignore either behavioral dynamics or context sparsity, and helps platform managers maximize long-term profit from in-session personalization.

Simulating CEO–Board Dynamics with Large Language Model (LLM) Agents Lynn Wu, Associate Professor, Operations, Information, and Decisions CEO–board dynamics shape some of the most consequential decisions of a firm. Whether a company discloses bad news, takes on risk, or pivots strategy often depends on the interplay between the CEO and the board.

Many empirical studies have documented relationships between CEO and board characteristics and firm outcomes. However, the underlying mechanisms remain underexplored, largely because large-scale data on CEO–board interactions are scarce.

This project seeks to develop an LLM-powered multi-agent simulation framework to examine how CEO–board interactions shape corporate strategy and to identify the mechanisms through which those interactions affect firm outcomes.

Balancing Cognitive Surrender and Offloading: Guardrails for Calibrated AI Use in Consumer Decisions Gideon Nave, Associate Professor, Marketing People increasingly rely on generative AI to answer questions and make decisions.

Shaw & Nave (2026) propose Tri‑System Theory of Cognition, extending dual‑process accounts by adding System 3 (artificial cognition) and documenting cognitive surrender: the tendency to adopt AI outputs with minimal scrutiny. This reliance can be adaptive when AI is correct, but harmful when AI is wrong.

This project extends Tri‑System Theory into consumer and business decision contexts, and tests when and how people shift between cognitive surrender (uncritical adoption) and cognitive offloading (strategic delegation with monitoring).

Across preregistered experiments, we will evaluate whether AI interface “guardrails”, including uncertainty signaling, lightweight verification prompts, and social accountability cues, can promote calibrated AI use: preserving performance gains when AI is reliable while increasing selective override when AI is unreliable.

The goal is to generate actionable guidance for consumers, marketers and practitioners who want to capture the benefits of AI assistance without eroding deliberative judgment or accountability in contexts where errors are costly.

Editing Digital Twins for Behavioral Fidelity Stefano Puntoni, Professor, Marketing Large language models (LLMs) have enabled “digital twins”—AI agents designed to simulate how specific consumers would respond to marketing questions and interventions.

While current digital twins can match average survey answers reasonably well, they often fail on what matters most for marketing: reproducing how people change their behavior under interventions such as anchors, sunk-cost cues, framing, and pricing changes. Instead, they behave too “assistant-like,” correcting biases and producing overly consistent, normative responses.

This project develops a new framework for intervention-consistent digital twins: models that mimic human responses not only on average, but under experimentally manipulated contexts, including realistic heterogeneity and noise.

Using the Twin-2K-500 benchmark dataset, we will combine conditional behavioral editing (activation/representation steering applied only in relevant experimental contexts), person-specific susceptibility to behavioral effects, and variance calibration grounded in human test–retest reliability. For pricing tasks, we will add lightweight economic constraints to ensure plausible demand responses.

The result will be a scalable method to build digital twins that are behaviorally credible and better suited for marketing research and decision support.

The Effects of Taxing Prescription Opioids on Drug Utilization and Provider Behavior Jackson Reimer, PhD Student Marissa King, Professor, Health Care Management The United States overdose epidemic has reached new heights with approximately 100,000 people dying from a drug overdose three of the last four years (Ahmad et al. , 2025).

Common policy solutions, such as prescription drug monitoring programs, have put the onus on prescribers to limit inappropriate opioid access. Market-oriented solutions to curb consumption, such as a tax, have been proposed by local, state, and federal legislators, though they have been historically less common.

This project fills this knowledge gap by examining the impact of a novel excise tax of prescription opioids for retail pharmacies in New York (NY) on opioid utilization, pharmacy growth, and population health. As of 2019, opioids priced less than $0. 50 per unit are taxed $0.

0025 per morphine milligram equivalent (MME) and more expensive drugs are taxed $0. 015 per MME. Although opioids are similar to other addictive goods, the welfare implications of such a tax on consumption are theoretically ambiguous.

First, it remains unclear ex-ante whether an opioid tax would effectively reduce consumption, as the out-of-pocket cost of prescription drugs is often limited to a fixed coinsurance payment. In effect, consumers are partially insulated from direct price increases. Second, the tax might reduce consumer wellbeing if it impacts both clinically appropriate and inappropriate prescribing.

Finally, even if inappropriate prescribing is reduced, consumers might substitute away from retail pharmacies to more expensive care settings such as outpatient clinics. Therefore, the impact of opioid taxation remains empirically ambiguous yet critical for policymakers considering similar legislation’s effects on the health care sector.

Teaching Statistics and Data Science from First Principles: A Data-Oriented, AI-Enabled Course Joseph Rudoler, PhD Student Abraham Wyner, Professor, Statistics and Data Science This course provides an introduction to statistics with an emphasis on building intuition about uncertainty in data analysis and decision-making.

The course is structured around thinking critically about data generating processes (i.e. the random, real-world phenomena that produce data) and how to model them computationally.

An essential aspect of this course is the assumption that students will use modern Large Language Models (LLMs) to accelerate their learning — this reduces the barrier to entry for basic data science programming and allows students to remain focused on important concepts in both data management and probabilistic thinking, instead of being bogged down by implementation details.

Tech-Washing: Strategic Technology Positioning and the Narrative-Innovation Gap Fernando Stein, PhD Student Winston Dou, Associate Professor, Finance Markets assign different valuations to different types of technology. AI firms trade at premiums to software firms, software firms to hardware firms, hardware firms to traditional industries.

These valuation hierarchies create incentives for firms to reposition themselves toward higher-valued technology segments through disclosure, regardless of whether their innovation portfolios support the repositioning. I term this practice “”tech-washing.

”” I document and measure tech-washing by constructing the first firm-year measure of the wedge between what companies say (technology narratives in SEC filings and earnings calls) and what they do (patented innovations). My measure combines natural language processing of corporate disclosure and news with patent portfolio analysis to identify firms whose technology positioning outpaces their technology substance.

The Effect of Biometric Payment Methods on Consumer Behavior Wendy De La Rosa, Assistant Professor, Marketing AI has shepherded new payment methods for consumers, such as biometric payment methods like facial recognition or palm vein scanning.

These methods are becoming increasingly prevalent, with large retail stores like WholeFoods adopting palm vein scanning across all of their US stores and restaurants like Steak ’n Shake rolling out facial recognition payment nationwide. Yet, despite the prevalence of these new payment methods, little is known about how they influence consumers’ behavior at the point of sale (POS), including spending, tipping, and donating.

Beyond deciding whether and how much to spend, consumers are often asked to make additional decisions at the POS, such as how much to tip or whether to donate to a charity. In this work, we investigate how emerging payment technologies influence these downstream financial behaviors.

Generative AI, Ideological Polarization, and the Transformation of Online Political Discourse Lynn Wu, Associate Professor, Operations, Information, and Decisions This project investigates how generative AI reshapes political discourse by focusing on a novel and counterintuitive pattern: initial evidence suggests that the adoption of large language models increases ideological polarization while simultaneously improving affective tone and civility.

Rather than escalating hostility, AI-assisted political expression appears to amplify ideological alignment in a calmer, less toxic register. The core objective of the project is to uncover the mechanisms driving this divergence.

By analyzing interaction-level dynamics—such as reply alignment, narrative reinforcement, and algorithmic sycophancy—the research seeks to explain how generative AI can polarize what people express politically while moderating how those views are communicated.

Use and Development of AI Companion Tools for Populations Under Stress Pinar Yildirim, Associate Professor, Marketing Rapid urbanization, educational transitions, and labor-market mobility increasingly expose individuals to acute social isolation and psychological stress, particularly in low-resource settings where access to mental-health services is limited.

Advances in generative artificial intelligence (GenAI) have enabled the creation of AI companions—always-available conversational agents capable of providing empathic conversations, reflection, and coping support. While these tools are widely adopted, there is little causal evidence on their welfare effects, their interaction with existing digital behaviors, or their implications for real-world social connection.

By providing the first causal evidence on AI companionship in vulnerable populations in the Global South, this