pith. sign in

arxiv: 2604.06047 · v2 · submitted 2026-04-07 · 💻 cs.CY

Algorithmic Monoculture and its Critics

Brian Hedden , Manish Raghavan This is my paper

Pith reviewed 2026-05-10 18:30 UTC · model grok-4.3

classification 💻 cs.CY
keywords algorithmic monoculturealgorithmic decision-makingsystemic exclusionagency and gaminginformation aggregationcritiques of algorithmshiring and lending algorithms
0
0 comments X

The pith

Algorithmic monoculture faces fewer decisive objections than its critics claim.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper takes on worries that a single algorithm deciding everything in hiring, lending, or justice will create systemic exclusion, erode agency through gaming, or block useful information diversity. It formalizes these and related objections, then tests them one by one. Many standard critiques turn out not to hold, while the remaining ones, though real, stop short of ruling out monoculture. The result is that the consistency promised by using one algorithm looks more defensible than the literature has assumed.

Core claim

We systematically evaluate a range of objections to monoculture, formalizing and rigorously assessing familiar critiques alongside novel ones. These objections concern systemic exclusion, agency and gaming, and information aggregation and exploration. We conclude that monoculture is less problematic than its critics have supposed: commonly cited objections fail, and while other objections have some force, they are not decisive against monoculture in general.

What carries the argument

Formalization of three main objection families (systemic exclusion, agency and gaming, information aggregation and exploration) followed by case-by-case assessment of whether each objection succeeds against monoculture.

If this is right

  • Consistency gains from monoculture can be pursued without automatically triggering systemic exclusion.
  • Objections based on strategic gaming or lost human agency do not rule out single-algorithm systems.
  • Concerns about reduced exploration or information loss are real but do not outweigh monoculture's advantages in every setting.
  • Policy arguments that demand algorithmic diversity lack decisive support from the standard critiques.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Regulators could safely permit more uniform algorithmic tools in high-stakes domains once the listed objections are addressed.
  • The same formal-evaluation approach could be applied to monoculture debates in other areas such as recommendation systems or medical diagnosis.
  • Empirical tests could measure whether real monoculture deployments produce the modest remaining harms the paper identifies.

Load-bearing premise

The formal versions of the critics' objections fully capture what those critics actually meant to argue.

What would settle it

An empirical case study of a domain that adopted full monoculture and then exhibited one of the specific harms the paper concludes are not decisive.

Figures

Figures reproduced from arXiv: 2604.06047 by Brian Hedden, Manish Raghavan.

Figure 1
Figure 1. Figure 1: Performance in Sequential Hiring Setting. [PITH_FULL_IMAGE:figures/full_fig_p019_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Performance in Simultaneous Hiring Setting. [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Failure to Identify Best Arm in Monoculture and Polyculture. [PITH_FULL_IMAGE:figures/full_fig_p024_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Total Bayesian Regret [PITH_FULL_IMAGE:figures/full_fig_p027_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Number of Arms Misclassified by Impartial Observer. [PITH_FULL_IMAGE:figures/full_fig_p027_5.png] view at source ↗
read the original abstract

Algorithmic decision-making is replacing idiosyncratic human judgment in domains such as hiring, lending, and criminal justice. This shift promises increased consistency, but many scholars worry that it can go too far. They warn of the dangers of algorithmic monoculture, in which all decisions across a domain are made using a single algorithm. We systematically evaluate a range of objections to monoculture, formalizing and rigorously assessing familiar critiques alongside novel ones. These objections concern systemic exclusion, agency and gaming, and information aggregation and exploration. We conclude that monoculture is less problematic than its critics have supposed: commonly cited objections fail, and while other objections have some force, they are not decisive against monoculture in general.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper claims that algorithmic monoculture—where a single algorithm handles decisions across an entire domain such as hiring or lending—is less problematic than critics have argued. It formalizes three families of objections (systemic exclusion, agency and gaming, and information aggregation and exploration), rigorously assesses them, and concludes that commonly cited objections fail while the remaining ones have limited force and are not decisive against monoculture in general.

Significance. If the formalizations and assessments are accurate, the paper offers a valuable conceptual clarification in AI ethics and algorithmic fairness. By distinguishing failed objections from those with partial merit, it provides a framework that could temper calls for mandatory algorithmic diversity and guide more targeted policy interventions. The work's strength lies in its systematic treatment of critiques rather than empirical claims or derivations.

major comments (2)
  1. [Abstract and Section on Formalizing Objections] The central conclusion that objections to monoculture 'fail' or 'are not decisive' depends on the specific formalizations of the three objection families. Without explicit verification that these formalizations preserve the full logical structure and empirical scope of the original critiques (particularly for systemic exclusion), there remains a risk that the rebuttals address only stylized versions rather than the strongest available arguments.
  2. [Section on Information Aggregation and Exploration] The assessment that information aggregation and exploration objections 'have some force' but are 'not decisive' requires a clear metric or threshold for decisiveness. The manuscript should specify what would count as decisive (e.g., a minimum expected loss in exploration or a particular failure mode in aggregation) to make the claim falsifiable and proportionate to the evidence presented.
minor comments (2)
  1. [Abstract] The abstract states the conclusions clearly but does not preview the structure of the formalizations; adding a brief roadmap sentence would improve readability for readers unfamiliar with the objection families.
  2. [Formalization sections] Notation for the formalized objections (e.g., variables representing exclusion rates or gaming payoffs) should be introduced consistently and defined on first use to avoid ambiguity in later sections.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which help clarify the scope and precision of our arguments. We address each major point below and indicate the revisions we will make.

read point-by-point responses

  1. Referee: [Abstract and Section on Formalizing Objections] The central conclusion that objections to monoculture 'fail' or 'are not decisive' depends on the specific formalizations of the three objection families. Without explicit verification that these formalizations preserve the full logical structure and empirical scope of the original critiques (particularly for systemic exclusion), there remains a risk that the rebuttals address only stylized versions rather than the strongest available arguments.

    Authors: We appreciate this concern and agree that explicit mapping strengthens the paper. Our formalizations were derived directly from the core logical structures in the cited literature (e.g., correlated errors for systemic exclusion, strategic manipulation for agency objections). To eliminate any ambiguity, we will add a dedicated subsection (new Section 3.4) that explicitly traces each formalized objection back to representative arguments from the primary sources, confirming that key empirical scope and logical elements are retained. This revision directly addresses the risk of stylized versions. revision: yes

  2. Referee: [Section on Information Aggregation and Exploration] The assessment that information aggregation and exploration objections 'have some force' but are 'not decisive' requires a clear metric or threshold for decisiveness. The manuscript should specify what would count as decisive (e.g., a minimum expected loss in exploration or a particular failure mode in aggregation) to make the claim falsifiable and proportionate to the evidence presented.

    Authors: We agree that greater precision on 'decisive' improves the analysis. In our framework, an objection is decisive against monoculture in general only if it shows that the costs of uniformity outweigh the benefits of consistency across a broad range of plausible parameter values and domains. We will revise the relevant section to articulate this criterion explicitly, including illustrative conditions (e.g., when the marginal value of exploration exceeds consistency gains by a factor that cannot be offset by other mechanisms). This renders the assessment more falsifiable while remaining consistent with the conceptual nature of the paper; we do not introduce a single numerical threshold because the appropriate cutoff is domain-dependent. revision: partial

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper conducts a conceptual and logical evaluation of objections to algorithmic monoculture by formalizing three families of critiques (systemic exclusion, agency/gaming, information aggregation/exploration) and assessing their force. No equations, derivations, fitted parameters, or first-principles predictions are present that could reduce by construction to the paper's own inputs. The analysis relies on direct reasoning about the stated objections rather than self-definitional structures, self-citation chains, or renaming of known results. The central claim—that monoculture is less problematic than critics suppose—emerges from this independent evaluation and remains self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim depends on domain assumptions about how objections can be formalized and assessed without quantitative parameters or new entities.

axioms (1)
  • domain assumption Objections to algorithmic monoculture can be formalized in a manner that permits rigorous logical assessment of their force.
    Invoked when the paper states it formalizes and evaluates the critiques.

pith-pipeline@v0.9.0 · 5402 in / 1170 out tokens · 69520 ms · 2026-05-10T18:30:32.549182+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

58 extracted references · 58 canonical work pages

  1. [1]

    Sample mean based index policies by O( n) regret for the multi-armed bandit problem

    Rajeev Agrawal. Sample mean based index policies by O( n) regret for the multi-armed bandit problem. Advances in applied probability, 27 0 (4): 0 1054--1078, 1995

    work page 1995

  2. [2]

    An auditing imperative for automated hiring systems

    Ifeoma Ajunwa. An auditing imperative for automated hiring systems. Harv. JL & Tech., 34: 0 621, 2020

    work page 2020

  3. [3]

    Machine bias

    Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner. Machine bias. ProPublica, May 2016. URL https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing

    work page 2016

  4. [4]

    Assigning more students to their top choices: A comparison of tie-breaking rules

    Itai Ashlagi, Afshin Nikzad, and Assaf Romm. Assigning more students to their top choices: A comparison of tie-breaking rules. Games and Economic Behavior, 115: 0 167--187, 2019

    work page 2019

  5. [5]

    Finite-time analysis of the multiarmed bandit problem

    Peter Auer, Nicolo Cesa-Bianchi, and Paul Fischer. Finite-time analysis of the multiarmed bandit problem. Machine learning, 47 0 (2): 0 235--256, 2002

    work page 2002

  6. [6]

    A supply and demand framework for two-sided matching markets

    Eduardo M Azevedo and Jacob D Leshno. A supply and demand framework for two-sided matching markets. Journal of Political Economy, 124 0 (5): 0 1235--1268, 2016

    work page 2016

  7. [7]

    Bandits with knapsacks

    Ashwinkumar Badanidiyuru, Robert Kleinberg, and Aleksandrs Slivkins. Bandits with knapsacks. Journal of the ACM (JACM), 65 0 (3): 0 1--55, 2018

    work page 2018

  8. [8]

    Strategic hiring under algorithmic monoculture

    Jackie Baek, Hamsa Bastani, and Shihan Chen. Strategic hiring under algorithmic monoculture. arXiv preprint arXiv:2502.20063, 2025

    work page arXiv 2025

  9. [9]

    The algorithm game

    Jane Bambauer and Tal Zarsky. The algorithm game. Notre Dame L. Rev., 94: 0 1, 2018

    work page 2018

  10. [10]

    Banerjee

    Abhijit V. Banerjee. A simple model of herd behavior. The quarterly journal of economics, 107 0 (3): 0 797--817, 1992

    work page 1992

  11. [11]

    Bandit social learning: Exploration under myopic behavior

    Kiarash Banihashem, Mohammad Taghi Hajiaghayi, Suho Shin, and Aleksandrs Slivkins. Bandit social learning: Exploration under myopic behavior. In Advances in Neural Information Processing Systems, 2023

    work page 2023

  12. [12]

    Big data's disparate impact

    Solon Barocas and Andrew D Selbst. Big data's disparate impact. Calif. L. Rev., 104: 0 671, 2016

    work page 2016

  13. [13]

    Mostly exploration-free algorithms for contextual bandits

    Hamsa Bastani, Mohsen Bayati, and Khashayar Khosravi. Mostly exploration-free algorithms for contextual bandits. Management Science, 67 0 (3): 0 1329--1349, 2021

    work page 2021

  14. [14]

    A theory of fads, fashion, custom, and cultural change as informational cascades

    Sushil Bikhchandani, David Hirshleifer, and Ivo Welch. A theory of fads, fashion, custom, and cultural change as informational cascades. Journal of political Economy, 100 0 (5): 0 992--1026, 1992

    work page 1992

  15. [15]

    The monoculture risk put into context

    Kenneth P Birman and Fred B Schneider. The monoculture risk put into context. IEEE Security & Privacy, 7 0 (1): 0 14--17, 2009

    work page 2009

  16. [16]

    Large deviations for martingales

    David Blackwell. Large deviations for martingales. In Festschrift for Lucien Le Cam: Research Papers in Probability and Statistics, pages 89--91. Springer, 1997

    work page 1997

  17. [17]

    Picking on the same person: Does algorithmic monoculture lead to outcome homogenization? Advances in Neural Information Processing Systems, 35: 0 3663--3678, 2022

    Rishi Bommasani, Kathleen A Creel, Ananya Kumar, Dan Jurafsky, and Percy S Liang. Picking on the same person: Does algorithmic monoculture lead to outcome homogenization? Advances in Neural Information Processing Systems, 35: 0 3663--3678, 2022

    work page 2022

  18. [18]

    Fairness

    John Broome. Fairness. Proceedings of the Aristotelian Society, 91 0 (1): 0 87--102, 1991. doi:10.1093/aristotelian/91.1.87

    work page doi:10.1093/aristotelian/91.1.87 1991

  19. [19]

    Wage bargaining with on-the-job search: Theory and evidence

    Pierre Cahuc, Fabien Postel-Vinay, and Jean-Marc Robin. Wage bargaining with on-the-job search: Theory and evidence. Econometrica, 74 0 (2): 0 323--364, 2006

    work page 2006

  20. [20]

    Filterworld: how algorithms flattened culture

    Kyle Chayka. Filterworld: how algorithms flattened culture. Random House, 2025

    work page 2025

  21. [21]

    The algorithmic leviathan: Arbitrariness, fairness, and opportunity in algorithmic decision-making systems

    Kathleen Creel and Deborah Hellman. The algorithmic leviathan: Arbitrariness, fairness, and opportunity in algorithmic decision-making systems. Canadian Journal of Philosophy, 52 0 (1): 0 26--43, 2022

    work page 2022

  22. [22]

    Kathleen A. Creel. Algorithmic monoculture and systemic exclusion. manuscript

    work page

  23. [23]

    Theory of statistical estimation

    Ronald Aylmer Fisher. Theory of statistical estimation. In Mathematical proceedings of the Cambridge philosophical society, volume 22, pages 700--725. Cambridge University Press, 1925

    work page 1925

  24. [24]

    College admissions and the stability of marriage

    David Gale and Lloyd S Shapley. College admissions and the stability of marriage. The American mathematical monthly, 69 0 (1): 0 9--15, 1962

    work page 1962

  25. [25]

    Goodin and Kai Spiekermann

    Robert E. Goodin and Kai Spiekermann. An Epistemic Theory of Democracy. Oxford University Press, Oxford, United Kingdom, 2018

    work page 2018

  26. [26]

    Personalized expertise search at linkedin

    Viet Ha-Thuc, Ganesh Venkataraman, Mario Rodriguez, Shakti Sinha, Senthil Sundaram, and Lin Guo. Personalized expertise search at linkedin. In 2015 IEEE International Conference on Big Data (Big Data), pages 1238--1247. IEEE, 2015

    work page 2015

  27. [27]

    Hamermesh and Jeff E

    Daniel S. Hamermesh and Jeff E. Biddle. Beauty and the labor market. The American Economic Review, 84 0 (5): 0 1174--1194, 1994. ISSN 00028282. URL http://www.jstor.org/stable/2117767

    work page arXiv 1994

  28. [28]

    Groups of diverse problem solvers can outperform groups of high-ability problem solvers

    Lu Hong and Scott E Page. Groups of diverse problem solvers can outperform groups of high-ability problem solvers. Proceedings of the National Academy of Sciences, 101 0 (46): 0 16385--16389, 2004

    work page 2004

  29. [29]

    Political skill: Explaining the effects of nonnative accent on managerial hiring and entrepreneurial investment decisions

    Laura Huang, Marcia Frideger, and Jone L Pearce. Political skill: Explaining the effects of nonnative accent on managerial hiring and entrepreneurial investment decisions. Journal of Applied Psychology, 98 0 (6): 0 1005, 2013

    work page 2013

  30. [30]

    Position: Scarce resource allocations that rely on machine learning should be randomized

    Shomik Jain, Kathleen Creel, and Ashia Wilson. Position: Scarce resource allocations that rely on machine learning should be randomized. In Proceedings of the 41st International Conference on Machine Learning, volume 235 of Proceedings of Machine Learning Research, pages 21148--21169. PMLR, 21--27 Jul 2024 a . URL https://proceedings.mlr.press/v235/jain24a.html

    work page 2024

  31. [31]

    Algorithmic pluralism: A structural approach to equal opportunity

    Shomik Jain, Vinith Suriyakumar, Kathleen Creel, and Ashia Wilson. Algorithmic pluralism: A structural approach to equal opportunity. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency, pages 197--206, 2024 b

    work page 2024

  32. [32]

    A smoothed analysis of the greedy algorithm for the linear contextual bandit problem

    Sampath Kannan, Jamie H Morgenstern, Aaron Roth, Bo Waggoner, and Zhiwei Steven Wu. A smoothed analysis of the greedy algorithm for the linear contextual bandit problem. Advances in Neural Information Processing Systems, 31, 2018

    work page 2018

  33. [33]

    The division of cognitive labor

    Philip Kitcher. The division of cognitive labor. Journal of Philosophy, 87 0 (1): 0 5--22, 1990. doi:10.2307/2026796

    work page doi:10.2307/2026796 1990

  34. [34]

    How do classifiers induce agents to invest effort strategically? ACM Transactions on Economics and Computation (TEAC), 8 0 (4): 0 1--23, 2020

    Jon Kleinberg and Manish Raghavan. How do classifiers induce agents to invest effort strategically? ACM Transactions on Economics and Computation (TEAC), 8 0 (4): 0 1--23, 2020

    work page 2020

  35. [35]

    Algorithmic monoculture and social welfare

    Jon Kleinberg and Manish Raghavan. Algorithmic monoculture and social welfare. Proceedings of the National Academy of Sciences, 118 0 (22): 0 e2018340118, 2021

    work page 2021

  36. [36]

    Algorithms as discrimination detectors

    Jon Kleinberg, Jens Ludwig, Sendhil Mullainathan, and Cass R Sunstein. Algorithms as discrimination detectors. Proceedings of the National Academy of Sciences, 117 0 (48): 0 30096--30100, 2020

    work page 2020

  37. [37]

    Price of anarchy of algorithmic monoculture

    Robert Kleinberg, Erald Sinanaj, and \'E va Tardos. Price of anarchy of algorithmic monoculture. In 21st Conference on Web and Internet Economics, 2025

    work page 2025

  38. [38]

    A. N. Kolmogorov. \:U ber das gesetz des iterierten loearithmus. Mathematische Annalen, 101: 0 126--135, 1929

    work page 1929

  39. [39]

    Worst-case equilibria

    Elias Koutsoupias and Christos Papadimitriou. Worst-case equilibria. In Annual symposium on theoretical aspects of computer science, pages 404--413. Springer, 1999

    work page 1999

  40. [40]

    Asymptotically efficient adaptive allocation rules

    Tze Leung Lai and Herbert Robbins. Asymptotically efficient adaptive allocation rules. Advances in applied mathematics, 6 0 (1): 0 4--22, 1985

    work page 1985

  41. [41]

    Hiring as exploration

    Danielle Li, Lindsey Raymond, and Peter Bergman. Hiring as exploration. Review of Economic Studies, page rdaf040, 2025

    work page 2025

  42. [42]

    The general theory of second best

    Richard G Lipsey and Kelvin Lancaster. The general theory of second best. The Review of Economic Studies, 24 0 (1): 0 11--32, 1956

    work page 1956

  43. [43]

    What kind of news gatekeepers do we want machines to be? Filter bubbles, fragmentation, and the normative dimensions of algorithmic recommendations

    Efrat Nechushtai and Seth C Lewis. What kind of news gatekeepers do we want machines to be? Filter bubbles, fragmentation, and the normative dimensions of algorithmic recommendations. Computers in human behavior, 90: 0 298--307, 2019

    work page 2019

  44. [44]

    Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy

    Cathy O'Neil. Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. Crown, New York, 2016

    work page 2016

  45. [45]

    Monoculture in matching markets

    Kenny Peng and Nikhil Garg. Monoculture in matching markets. Advances in Neural Information Processing Systems, 37: 0 81959--81991, 2024 a

    work page 2024

  46. [46]

    Wisdom and foolishness of noisy matching markets

    Kenny Peng and Nikhil Garg. Wisdom and foolishness of noisy matching markets. In Proceedings of the 25th ACM Conference on Economics and Computation, page 675, 2024 b

    work page 2024

  47. [47]

    Competition and diversity in generative AI

    Manish Raghavan. Competition and diversity in generative AI . arXiv preprint arXiv:2412.08610, 2024

    work page arXiv 2024

  48. [48]

    Greedy algorithm almost dominates in smoothed contextual bandits

    Manish Raghavan, Aleksandrs Slivkins, Jennifer Wortman Vaughan, and Zhiwei Steven Wu. Greedy algorithm almost dominates in smoothed contextual bandits. SIAM Journal on Computing, 52 0 (2): 0 487--524, 2023

    work page 2023

  49. [49]

    LLM reality check: The hidden instability of ai r \'e sum \'e screening

    Martyn Redstone. LLM reality check: The hidden instability of ai r \'e sum \'e screening. Technical report, Eunomia HR, June 2025

    work page 2025

  50. [50]

    Experimental study of inequality and unpredictability in an artificial cultural market

    Matthew J Salganik, Peter Sheridan Dodds, and Duncan J Watts. Experimental study of inequality and unpredictability in an artificial cultural market. Science, 311 0 (5762): 0 854--856, 2006

    work page 2006

  51. [51]

    What We Owe to Each Other

    Thomas Scanlon. What We Owe to Each Other. Harvard University Press, Cambridge, 1998

    work page 1998

  52. [52]

    The measurement of mobility

    Anthony F Shorrocks. The measurement of mobility. Econometrica: Journal of the Econometric Society, pages 1013--1024, 1978

    work page 1978

  53. [53]

    Introduction to multi-armed bandits

    Aleksandrs Slivkins. Introduction to multi-armed bandits. Foundations and Trends in Machine Learning, 12 0 (1-2): 0 1--286, 2019

    work page 2019

  54. [54]

    Serendipity in the city: User evaluations of urban recommender systems

    Annelien Smets, Jorre Vannieuwenhuyze, and Pieter Ballon. Serendipity in the city: User evaluations of urban recommender systems. Journal of the Association for Information Science and Technology, 73 0 (1): 0 19--30, 2022

    work page 2022

  55. [55]

    Foreign accents and employer beliefs: Experimental evidence on hiring discrimination

    Elisa Taveras Pena, Ozlem Tonguc, Maria Zhu, and Nicola Miller. Foreign accents and employer beliefs: Experimental evidence on hiring discrimination. IZA Discussion Paper No. 18239, 2025

    work page 2025

  56. [56]

    Ecosystem-level analysis of deployed machine learning reveals homogeneous outcomes

    Connor Toups, Rishi Bommasani, Kathleen Creel, Sarah Bana, Dan Jurafsky, and Percy S Liang. Ecosystem-level analysis of deployed machine learning reveals homogeneous outcomes. Advances in Neural Information Processing Systems, 36: 0 51178--51201, 2023

    work page 2023

  57. [57]

    Judgment under uncertainty: Heuristics and biases: Biases in judgments reveal some heuristics of thinking under uncertainty

    Amos Tversky and Daniel Kahneman. Judgment under uncertainty: Heuristics and biases: Biases in judgments reveal some heuristics of thinking under uncertainty. Science, 185 0 (4157): 0 1124--1131, 1974

    work page 1974

  58. [58]

    The epistemic benefit of transient diversity

    Kevin JS Zollman. The epistemic benefit of transient diversity. Erkenntnis, 72 0 (1): 0 17--35, 2010

    work page 2010