







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. In this paper, we develop 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 individuals to directly manipulate algorithm parameters and visualize the resulting tradeoffs between competing fairness metrics. Across two preregistered online experiments (n = 1, 563) in a healthcare context, we demonstrate that lay participants consistently identify efficient, near-optimal trade-offs that approach the Pareto frontier. Yet, their specific preferences are fragile: they shift significantly based on seemingly innocuous details of the information design, such as the order in which metrics are presented. These findings show that (1) our framework can effectively reveal fairness preferences and (2) the design of the elicitation process itself can fundamentally shape how fairness is perceived.
Underspecified Human Decision Experiments Considered Harmful
Decision-making with information displays is a key focus of research in areas like human-AI collaboration and data visualization. However, what constitutes a decision problem, and what is required for an experiment to conclude that decisions are flawed, remain imprecise. We present a widely applicable definition of a decision problem synthesized from statistical decision theory and information economics. We claim that to attribute loss in human performance to bias, an experiment must provide the information that a rational agent would need to identify the normative decision. We evaluate whether recent empirical research on AI-assisted decisions achieves this standard. We find that only 10 (26%) of 39 studies that claim to identify biased behavior presented participants with sufficient information to make this claim in at least one treatment condition. We motivate the value of studying well-defined decision problems by describing a characterization of performance losses they allow to be conceived.

A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experiments
Environments built for people are increasingly operated by a new class of economic actors: LLM-powered software agents making decisions on our behalf. These decisions range from our purchases to travel plans to medical treatment selection. Current evaluations of these agents largely focus on task competence, but we argue for a deeper assessment: how these agents choose when faced with realistic decisions. We introduce ABxLab, a framework for systematically probing agentic choice through controlled manipulations of option attributes and persuasive cues. We apply this to a realistic web-based shopping environment, where we vary prices, ratings, and psychological nudges, all of which are factors long known to shape human choice. We find that agent decisions shift predictably and substantially in response, revealing that agents are strongly biased choosers even without being subject to the cognitive constraints that shape human biases. This susceptibility reveals both risk and opportunity: risk, because agentic consumers may inherit and amplify human biases; opportunity, because consumer choice provides a powerful testbed for a behavioral science of AI agents, just as it has for the study of human behavior. We release our framework as an open benchmark for rigorous, scalable evaluation of agent decision-making.

Characterizing Fairness Over the Set of Good Models Under Selective Labels
Algorithmic risk assessments are used to inform decisions in a wide variety of high-stakes settings. Often multiple predictive models deliver similar overall performance but differ markedly in their predictions for individual cases, an empirical phenomenon known as the “Rashomon Effect.” These models may have different properties over various groups, and therefore have different predictive fairness properties. We develop a framework for characterizing predictive fairness properties over the set of models that deliver similar overall performance, or “the set of good models.” Our framework addresses the empirically relevant challenge of selectively labelled data in the setting where the selection decision and outcome are unconfounded given the observed data features. Our framework can be used to 1) audit for predictive bias; or 2) replace an existing model with one that has better fairness properties. We illustrate these use cases on a recidivism prediction task and a real-world credit-scoring task.
Can Revealed Preferences Clarify LLM Alignment and Steering?
LLMs are increasingly used to make or support high-stakes decisions under uncertainty, where alignment depends not only on factual accuracy but on how models weigh tradeoffs between different outcomes. We present an empirical pipeline for estimating the implied preferences that an LLM's observed choices optimize: we elicit the model's probability distribution over unknowns along with the choice it would make for the decision task and then fit a discrete choice model to recover the cost function that best rationalizes the model's decisions. We show how this revealed-preference description allows rigorous evaluation of whether models behave in a consistently goal-directed way, whether they can verbalize a description of their objectives which matches their revealed decision policy, and whether prompting can reliably steer those policies to implement a user-specified cost function. We apply this evaluation across four medical diagnosis domains and multiple frontier and open-source models. We find that while many models have a nontrivial degree of internal coherence, they also have significant weaknesses in faithfully reporting or adopting preferences in response to user direction.

Picking on the Same Person: Does Algorithmic Monoculture lead to Outcome Homogenization?
As the scope of machine learning broadens, we observe a recurring theme of algorithmic monoculture: the same systems, or systems that share components (e.g. datasets, models), are deployed by multiple decision-makers. While sharing offers advantages like amortizing effort, it also has risks. We introduce and formalize one such risk, outcome homogenization: the extent to which particular individuals or groups experience the same outcomes across different deployments. If the same individuals or groups exclusively experience undesirable outcomes, this may institutionalize systemic exclusion and reinscribe social hierarchy. We relate algorithmic monoculture and outcome homogenization by proposing the component sharing hypothesis: if algorithmic systems are increasingly built on the same data or models, then they will increasingly homogenize outcomes. We test this hypothesis on algorithmic fairness benchmarks, demonstrating that increased data-sharing reliably exacerbates homogenization and individual-level effects generally exceed group-level effects. Further, given the current regime in AI of foundation models, i.e. pretrained models that can be adapted to myriad downstream tasks, we test whether model-sharing homogenizes outcomes across tasks. We observe mixed results: we find that for both vision and language settings, the specific methods for adapting a foundation model significantly influence the degree of outcome homogenization. We also identify societal challenges that inhibit the measurement, diagnosis, and rectification of outcome homogenization in deployed machine learning systems.
The Challenge of Understanding What Users Want: Inconsistent Preferences and Engagement Optimization
Online platforms have a wealth of data, run countless experiments, and use industrial-scale algorithms to optimize user experience. Despite this, many users seem to regret the time they spend on these platforms. One possible explanation is that incentives are misaligned: platforms are not optimizing for user happiness. We suggest the problem runs deeper, transcending the specific incentives of any particular platform, and instead stems from a mistaken foundational assumption. To understand what users want, platforms look at what users do. This is a kind of revealed-preference assumption that is ubiquitous in the way user models are built. Yet research has demonstrated, and personal experience affirms, that we often make choices in the moment that are inconsistent with what we actually want. The behavioral economics and psychology literatures suggest, for example, that we can choose mindlessly or that we can be too myopic in our choices, behaviors that feel entirely familiar on online platforms. In this work, we develop a model of media consumption where users have inconsistent preferences. We consider a platform which wants to maximize user utility, but only observes behavioral data in the form of the user’s engagement. We show how our model of users’ preference inconsistencies produces phenomena that are familiar from everyday experience but difficult to capture in traditional user interaction models. These phenomena include users who have long sessions on a platform but derive very little utility from it, and platform changes that steadily raise user engagement before abruptly causing users to go “cold turkey” and quit. A key ingredient in our model is a formulation for how platforms determine what to show users: they optimize over a large set of potential content (the content manifold) parametrized by underlying features of the content. Whether improving engagement improves user welfare depends on the direction of movement in the content manifold: For certain directions of change, increasing engagement makes users less happy, whereas in other directions on the same manifold, increasing engagement makes users happier. We provide a characterization of the structure of content manifolds for which increasing engagement fails to increase user utility. By linking these effects to abstractions of platform design choices, our model thus creates a theoretical framework and vocabulary in which to explore interactions between design, behavioral science, and social media. This paper was accepted by Yan Chen, behavioral economics and decision analysis. Funding: This work was supported by the Vannevar Bush Faculty Fellowship and Multidisciplinary University Research Initiative [Grant W911NF-19-0217]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/mnsc.2022.03683 .

Can Consumers Make Affordable Care Affordable? The Value of Choice Architecture
Tens of millions of people are currently choosing health coverage on a state or federal health insurance exchange as part of the Patient Protection and Affordable Care Act. We examine how well people make these choices, how well they think they do, and what can be done to improve these choices. We conducted 6 experiments asking people to choose the most cost-effective policy using websites modeled on current exchanges. Our results suggest there is significant room for improvement. Without interventions, respondents perform at near chance levels and show a significant bias, overweighting out-of-pocket expenses and deductibles. Financial incentives do not improve performance, and decision-makers do not realize that they are performing poorly. However, performance can be improved quite markedly by providing calculation aids, and by choosing a “smart” default. Implementing these psychologically based principles could save purchasers of policies and taxpayers approximately 10 billion dollars every year.
Medical Algorithms Are Failing Communities Of Color
Medical algorithms routinely make decisions about patients’ health care, yet they are rife with bias. Health equity must be built into the development and deployment of these health algorithms.
Task-Dependent Algorithm Aversion
Research suggests that consumers are averse to relying on algorithms to perform tasks that are typically done by humans, despite the fact that algorithms often perform better. The authors explore when and why this is true in a wide variety of domains. They find that algorithms are trusted and relied on less for tasks that seem subjective (vs. objective) in nature. However, they show that perceived task objectivity is malleable and that increasing a task’s perceived objectivity increases trust in and use of algorithms for that task. Consumers mistakenly believe that algorithms lack the abilities required to perform subjective tasks. Increasing algorithms’ perceived affective human-likeness is therefore effective at increasing the use of algorithms for subjective tasks. These findings are supported by the results of four online lab studies with over 1,400 participants and two online field studies with over 56,000 participants. The results provide insights into when and why consumers are likely to use algorithms and how marketers can increase their use when they outperform humans.

Task-Dependent Algorithm Aversion
Research suggests that consumers are averse to relying on algorithms to perform tasks that are typically done by humans, despite the fact that algorithms often perform better. The authors explore when and why this is true in a wide variety of domains. They find that algorithms are trusted and relied on less for tasks that seem subjective (vs. objective) in nature. However, they show that perceived task objectivity is malleable and that increasing a task’s perceived objectivity increases trust in and use of algorithms for that task. Consumers mistakenly believe that algorithms lack the abilities required to perform subjective tasks. Increasing algorithms’ perceived affective human-likeness is therefore effective at increasing the use of algorithms for subjective tasks. These findings are supported by the results of four online lab studies with over 1,400 participants and two online field studies with over 56,000 participants. The results provide insights into when and why consumers are likely to use algorithms and how marketers can increase their use when they outperform humans.

Human-Algorithm Collaboration with Private Information: Naïve Advice-Weighting Behavior and Mitigation
Even if algorithms make better predictions than humans on average, humans may sometimes have private information that an algorithm does not have access to that can improve performance. How can we help humans effectively use and adjust recommendations made by algorithms in such situations? When deciding whether and how to override an algorithm’s recommendations, we hypothesize that people are biased toward following naïve advice-weighting (NAW) behavior; they take a weighted average between their own prediction and the algorithm’s prediction, with a constant weight across prediction instances regardless of whether they have valuable private information. This leads to humans overadhering to the algorithm’s predictions when their private information is valuable and underadhering when it is not. In an online experiment where participants were tasked with making demand predictions for 20 products while having access to an algorithm’s predictions, we confirm this bias toward NAW and find that it leads to a 20%–61% increase in prediction error. In a second experiment, we find that feature transparency—even when the underlying algorithm is a black box—helps users more effectively discriminate how to deviate from algorithms, resulting in a 25% reduction in prediction error. We make further improvements in a third experiment via an intervention designed to move users away from advice weighting and instead, use only their private information to inform deviations, leading to a 34% reduction in prediction error. This paper was accepted by Elena Katok, Special Issue on the Human-Algorithm Connection. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.03850 .

Preference Filtering: When Customers Share Narrow Preferences with Algorithms
Digital platforms commonly ask customers to select interest categories (e.g., genres/topics) as input for personalized recommendations. Twelve main studies and two pilot studies (total n = 8,808) reveal that customers share less diverse preferences with algorithms (vs. human curators or when merely listing preferences for themselves); they focus on core preferences while omitting tangential ones, a phenomenon termed “preference filtering.” It is driven by customers’ expectation that algorithms weight their preferences more uniformly than human curators (i.e., expected uniformity). A mathematical model identifies conditions under which preference filtering appears rational ex ante yet reduces recommendation fit ex post. Interviews and a survey suggest that customers expect more uniform weighting than practitioners do. The authors examine key design dimensions of the preference elicitation task—when preferences are elicited, how customers articulate them, what purpose is made salient, and who customers believe they are interacting with—that determine the extent to which customers engage in preference filtering. Two studies on self-developed video-streaming websites show that alleviating preference filtering can boost engagement and enhance customer reviews of recommendation services. These findings offer valuable insights for firms that rely on algorithms to engage customers.

The Effects of Search Interface on Consumer Preferences: Evidence from a Randomized Experiment
While structured search interfaces—such as filters and sorting tools—help consumers reduce search frictions on digital platforms, they also shape how consumers encounter product information during search. This paper studies how search interface design influences revealed preferences and market outcomes. We develop a theoretical framework linking interface design to attribute preferences through salience reallocation and information-based preference updating. The framework generates two predictions that can be formally derived: restricting faceted filters reduces the concentration of attribute importance weights (i.e., preference dispersion) and increases the diversity of products chosen (i.e., choice diversity). We test these predictions using a randomized, incentive-aligned online experiment in which 1,200 participants searched for wireless headphones on a custom-built e-commerce platform. Restricting filters led consumers to explore more broadly and resulted in more evenly distributed attribute importance weights and more diverse, higher-priced product choices. Mediation analysis further suggests that increased exploration was associated with shifts in attribute preferences. Our findings demonstrate that search interface design can construct, rather than merely reveal, consumer preferences.

People Reject Algorithms in Uncertain Decision Domains Because They Have Diminishing Sensitivity to Forecasting Error
Will people use self-driving cars, virtual doctors, and other algorithmic decision-makers if they outperform humans? The answer depends on the uncertainty inherent in the decision domain. We propose that people have diminishing sensitivity to forecasting error and that this preference results in people favoring riskier (and often worse-performing) decision-making methods, such as human judgment, in inherently uncertain domains. In nine studies ( N = 4,820), we found that (a) people have diminishing sensitivity to each marginal unit of error that a forecast produces, (b) people are less likely to use the best possible algorithm in decision domains that are more unpredictable, (c) people choose between decision-making methods on the basis of the perceived likelihood of those methods producing a near-perfect answer, and (d) people prefer methods that exhibit higher variance in performance (all else being equal). To the extent that investing, medical decision-making, and other domains are inherently uncertain, people may be unwilling to use even the best possible algorithm in those domains.

Certifying and Removing Disparate Impact
What does it mean for an algorithm to be biased? In U.S. law, unintentional bias is encoded via disparate impact, which occurs when a selection process has widely different outcomes for different groups, even as it appears to be neutral. This legal determination hinges on a definition of a protected class (ethnicity, gender) and an explicit description of the process.When computers are involved, determining disparate impact (and hence bias) is harder. It might not be possible to disclose the process. In addition, even if the process is open, it might be hard to elucidate in a legal setting how the algorithm makes its decisions. Instead of requiring access to the process, we propose making inferences based on the data it uses.We present four contributions. First, we link disparate impact to a measure of classification accuracy that while known, has received relatively little attention. Second, we propose a test for disparate impact based on how well the protected class can be predicted from the other attributes. Third, we describe methods by which data might be made unbiased. Finally, we present empirical evidence supporting the effectiveness of our test for disparate impact and our approach for both masking bias and preserving relevant information in the data. Interestingly, our approach resembles some actual selection practices that have recently received legal scrutiny.

Learn More - The Algorithmic Justice League
"We aim to empower communities and galvanize decision makers to take action that mitigates AI bias and harms."
