Pith. sign in

REVIEW 4 major objections 4 minor 101 references

Trustworthiness in Stochastic Systems: Towards Opening the Black Box

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read A stochastic AI system is trustworthy, this paper argues, exactly when its values align with the user's relevant values for the task, regardless of output randomness.

desk verdict A genuinely useful conceptual cleanup of stochasticity in AI trust, with a near-tautological central claim and an unoperationalized positive proposal. read the letter →

arxiv 2501.16461 v1 pith:JB2QHS34 submitted 2025-01-27 cs.CY

classification cs.CY
keywords trustworthinessstochasticityvaluealignmentlatentmodelingsociotechnicalsystemsgenerativeAItrustcausaldiagrams
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that randomness in an AI system is not, by itself, a threat to trustworthiness. The authors' central claim is that if the values embodied by the system align with the user's relevant values for a given task, the system is trustworthy regardless of how stochastic its outputs are. They show why two prevailing responses fail: removing all user-facing randomness can destroy valuable diversity and create a false sense of certainty, while letting users dial a single 'variability' control conflates distinct forms of stochasticity that matter to different values. In place of these, they propose a refined definition of stochasticity relative to a user's level of description and knowledge, and a latent value modeling framework for inferring the values of both system and user from a causal diagram of the pipeline. The payoff would be trust calibration based on value alignment rather than on output stability.

What carries the argument

The load-bearing machinery is a pair of causal diagrams—one from the user's perspective, one from the LLM's—joined into a single causal chain from prompt to output. Observed nodes (yellow) are things like prompts and outputs; red and blue nodes are latent states that the paper proposes to model: user-determined values (goal, prompt, interpretation) and system-determined values (developer guardrails, default prompt engineering, data, intermediate representations). The defining move is to treat these latent values as unobserved variables and estimate them with standard latent variable procedures such as expectation-maximization or Markov chain Monte Carlo, so that alignment can be evaluated from the inferred value distributions rather than from raw output variance. This is what the paper calls latent value modeling.

What would settle it

Run a controlled image-generation study where a system produces outputs that all satisfy the user's stated values but vary along dimensions the user marks as irrelevant, and compare trust ratings against a deterministic version with the same value-relevant behavior; if users' trust drops when value-irrelevant dimensions vary, the claim that value alignment is sufficient for trustworthiness fails.

Watch

Extended reading notes

Core claim

Section 5 states the paper's central claim directly: 'if the system's and user's relevant values are aligned for a given task, then the system is trustworthy, regardless of stochasticity.' The paper arrives at this by refining the notion of stochasticity: a system should count as stochastic for trust purposes only when its output variability occurs at or above the user's level of relevant description, or outside the user's knowledge—not merely because probabilistic processes exist somewhere in its pipeline. It then argues that trustworthiness is a normative, context-relative property of the trustee: B is trustworthy for A when value alignment is such that A should trust B. On this basis, the paper rejects both deterministic presentation and user-controlled stochasticity dials as inadequate, and proposes latent value modeling as a sociotechnical alternative that opens the black box by explicitly representing values at each causal stage of the system and of the user. The intended result is a framework for determining when stochastic variability actually undermines trust and when it is value-irrelevant noise.

Load-bearing premise

The framework rests on the assumption that the values embodied by an AI system and by a user can be represented as latent states and reliably inferred from observable outputs using standard latent variable estimation; if those states are not identifiable from what can be observed, the proposed trust assessment cannot be computed.

Editorial extensions

If this is right

  • Trust certification for stochastic AI should be expressed relative to a user's value profile and a task, not as a global reliability score based on output variance.
  • Eliminating all user-facing stochasticity is not a general fix: it can suppress value-relevant diversity and create an illusion that the fixed output is the only or correct one.
  • Single-dial stochasticity controls are inherently inadequate because they treat variation in input interpretation, content generation, and output selection as interchangeable.
  • Auditing a stochastic system becomes a value-inference problem: determine whether latent user and system values align, rather than measuring how often outputs repeat or fall within tolerance.
  • Behavioral alignment methods like RLHF remain incomplete because they can only capture values users can express through feedback, leaving implicit or unarticulated values unmodeled.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension: in a controlled study with a generative image model, hold overall output variance fixed but move it between dimensions users rate as value-relevant and value-irrelevant; the framework predicts trust judgments will track only the value-relevant variance.
  • The latent value inference step stands or falls on identifiability: the manuscript itself notes the space of latent variables must roughly resemble human values, so a natural extension is to determine empirically which causal diagrams yield unique value decompositions from outputs.
  • If the alignment criterion is right, global 'trustworthiness scores' are conceptually misplaced; trust metrics would need to be indexed by user values, task, and knowledge, making regulatory certification a matter of matching value profiles rather than engineering stability.
  • By analogy to the paper's appendix on interpersonal trust, adding signaling, explanation, and accountability mechanisms to AI systems could reduce the trust-eroding effect of stochasticity even when the underlying randomness is unchanged—a design direction the paper only gestures at.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. This paper argues that stochasticity in AI systems does not uniformly undermine trustworthiness; rather, it matters only when variability interferes with value alignment between the system and the user. The authors criticize two practical responses to stochasticity—eliminating user-facing variability and giving users control dials over variability—and then propose a refined definition of stochasticity as variation at or above the user's level of relevant description or beyond the user's knowledge. They further propose a 'latent value modeling' framework in which user and system values are treated as latent variables in causal diagrams, to be estimated from observable behavior and used to assess value alignment.

Significance. The paper's negative arguments are clear and useful: Section 4 convincingly identifies limitations of both deterministic-output strategies and user-controlled variability dials, and Section 3 gives a concrete account of how intermittent misalignment complicates trust formation. The paper is also honest about the speculative status of its positive proposal, and the illustrative running example of image generation aids readability. However, the central positive claim is currently close to tautological, and the proposed latent value modeling framework is underdeveloped: it lacks a formal model, identifiability analysis, and empirical instantiation. If the conceptual clarification and formalization were supplied, the framework could offer a valuable reframing of trust assessment in stochastic systems, but in its present form the contribution is primarily critical rather than constructive.

major comments (4)
  1. [Section 5] The central claim that 'if the system's and user's relevant values are aligned for a given task, then the system is trustworthy, regardless of stochasticity' is true by stipulation given that Section 2.3 defines trustworthiness as appropriate value alignment. The paper needs to give an independent semantics for value alignment that specifies whether alignment is evaluated ex ante over the output distribution or ex post on each output. Under an ex ante reading, a system that is safe 99% of the time and catastrophically misaligned 1% of the time could count as aligned, contradicting the Section 3.1 argument that intermittent misalignment undermines trustworthiness. Under an ex post reading, stochasticity is not irrelevant, because the probability of value-relevant misalignment becomes exactly what determines trustworthiness. The paper must resolve this equivocation before the 'regardless of stochasticity' claim can be assessed.
  2. [Section 5.1] The refined definition of a stochastic system—variability at or above the user-specific level of relevant description, or beyond the user's knowledge—presupposes a level of description for values but gives no account of how to aggregate over stochastic draws. For a user whose relevant value is safe operation, a 1% failure rate is either a violation at the relevant level of description (so alignment fails and stochasticity matters) or not a violation (so the system is declared trustworthy despite a known catastrophic risk). The paper needs a principled account of how probabilities of value-relevant outcomes are evaluated at the user's level of description; otherwise the definition can be used to classify any undesired variation as irrelevant.
  3. [Section 6.2] The claim that the red and blue nodes can be treated as 'mostly unobserved variables' and estimated using 'any number of latent variable estimation procedures' is not supported by a formal model. The paper does not specify the structural equations, the measurement model linking latent values to observed prompts and outputs, the identifiability conditions for the latent states, or the data requirements for estimation. This matters because Section 3.2.1 itself argues that behavioral distributions conditional on values are 'extremely complex, if not impossible to specify'; Section 6.2 does not explain how latent value modeling overcomes that difficulty. As written, the proposed trustworthiness assessment cannot be computed or empirically validated.
  4. [Section 6.1] The causal diagrams in Figures 1 and 2 are presented as 'intentionally simplified' and omit multiple connections, including between cultural and social norms and guardrails. Without a precise specification of which nodes and edges are included, and under what causal assumptions the graphs are valid, it is unclear what inferential query the model is intended to answer (for example, whether it supports counterfactual judgments about trustworthiness under hypothetical value changes). The paper should either state explicitly that the diagrams are purely illustrative and not yet a formal model, or provide the causal assumptions needed to make the proposed inference well-defined.
minor comments (4)
  1. [Section 6.1] The sentence 'By understand the latent values contributing to both perspectives' should read 'By understanding the latent values contributing to both perspectives'.
  2. [Figures 1 and 2] The manuscript references Figures 1 and 2, but the figures are not included in the submitted text; they need to be added so that the causal diagrams can be checked against the description in Section 6.1.
  3. [Section 5] The phrase 'before delving into two current approaches' is ambiguous because the following subsections discuss direct implementation, behavioral RLHF-based alignment, and then latent value modeling as a third approach; consider rewording to clarify the intended enumeration.
  4. [Section 3.2.1] The paper mentions that a posterior distribution could be used to calculate the probability that the system's behavior will fail to support user values, but this quantitative thread is not picked up later; connecting it to the latent value modeling proposal in Section 6 would strengthen the argument.

Circularity Check

2 steps flagged · score 8.0 of 10

Central claim is definitional: 'value alignment' is the same relation used to define 'trustworthiness,' so Section 5's conclusion is an unpacking of the definition, and the refined stochasticity definition builds the conclusion into its terms.

  1. self definitional [Section 2.3 (Trustworthiness) and Section 5 (Proposal: Value Alignment for Trustworthiness)]
    "In particular, B is trustworthy for A when there is appropriate value alignment such that A should trust B. ... we propose value alignment as a framework for addressing this challenge: if the system's and user's relevant values are aligned for a given task, then the system is trustworthy, regardless of stochasticity."

    The Section 5 antecedent ('relevant values are aligned') is the same relation used in Section 2.3 to define 'trustworthy': B is trustworthy for A when there is appropriate value alignment. The proposal is therefore not an independent derivation but an unpacking of the earlier definition. No independent semantics is supplied for 'appropriate value alignment' that would let one verify the conditional without already knowing trustworthiness, and the 'regardless of stochasticity' clause adds no constraint because any stochasticity that undermines alignment is reclassified as value-relevant under Section 5.1. The central thesis is thus true by definition rather than by argument.

  2. self definitional [Section 5.1 (Stochasticity (Refined)) and Section 6.1 (Latent Value Modeling)]
    "Specifically, we define a stochastic system as one that produces outputs (given a fixed input) that exhibit variability (i) at or above the user-specific level of relevant description (as determined by their values and input), or (ii) beyond their knowledge. ... The system's stochastic nature becomes a barrier to trust only when variability negatively impacts user values."

    The paper's recurring claim that 'not all forms of stochasticity affect trustworthiness in the same way or to the same degree' is built into the refined definition: a system counts as stochastic only when its variability occurs at or above the user-specific level of relevant description. Variation below that level is definitionally excluded, so the conclusion that only value-relevant variability threatens trust follows by construction. This is a stipulated definition, not an empirical or theoretical result, and it cannot independently support the main thesis.

full rationale

The paper's central result—that value alignment suffices for trustworthiness regardless of stochasticity—is imposed by the Section 2.3 definition of trustworthiness as appropriate value alignment. Section 5 re-states that definition as a conclusion, and Section 5.1's refined definition of stochasticity similarly encodes the key claim about which stochasticity matters. These are transparent stipulative moves, but they are still cases where the announced conclusion is equivalent to the input definition by construction. The paper contains no equation-level fitting, no machine-checked derivations, and no external benchmark against which the central claim is tested. The two self-citations (refs [7] and [70]) are not load-bearing, so they do not contribute to the score. The score of 8 reflects that the primary thesis is forced by definition rather than by independent evidence or argument; the material on critique of existing approaches and latent value modeling has independent descriptive content, which prevents the score from reaching 10.

Assumptions & free parameters 0 free parameters · 5 assumptions · 2 invented entities

No numeric free parameters are fitted. The paper's load-bearing assumptions are conceptual: the value-based definition of trust, the user-relative definition of stochasticity, and the assumption that latent values can be inferred from causal structure and behavior. The proposed latent value nodes are invented conceptual entities with no independent evidence, which is a significant gap for the positive framework.

assumptions (5)
  • domain assumption Trust requires justified expectations of value alignment and vulnerability (Section 2.1).
    The entire analysis adopts a normative, value-based conception of trust rather than a purely behavioral or predictive one.
  • domain assumption Trustworthiness is trustee-relative: B is trustworthy for A only when B supports A's relevant values (Section 2.3).
    This grounds the claim that stochasticity matters only when it affects user values.
  • domain assumption A system is stochastic if it has probabilistic components, and trust-relevant stochasticity is variation at or above the user's level of relevant description or beyond their knowledge (Sections 2.4, 5.1).
    The refined definition is stipulative and drives the later argument about which stochasticity threatens trust.
  • domain assumption Human values are relatively stable over time and inferable enough for alignment assessment (Appendix A).
    The paper argues that human trust works because values are stable and signaled, and it uses this to motivate the feasibility of latent value modeling.
  • ad hoc to paper Values can be treated as latent variables in a causal model and estimated from behavior (Section 6.2).
    This is the core enabling assumption of the proposed framework and is not independently established; the paper only gestures at latent variable estimation procedures.
invented entities (2)
  • User-determined latent value states (red nodes in causal diagrams)
    purpose: Represent unobserved user goals, values, and interpretations that drive prompting and output evaluation.
    Introduced in Section 6.1 and Figure 1; no operational measurement, no falsifiable prediction, and no independent evidence are provided.
  • LLM-determined latent value states (blue nodes in causal diagrams)
    purpose: Represent values embedded in training data, guardrails, prompt engineering, and intermediate representations.
    Introduced as conceptual nodes in Section 6.1; no estimation procedure, validation, or empirical handle is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Trustworthiness in Stochastic Systems: Towards Opening the Black Box." pith.science (2026). https://pith.science/paper/JB2QHS34

@misc{pith2026250116461,
  author       = {Pith},
  title        = {Pith review of: Trustworthiness in Stochastic Systems: Towards Opening the Black Box},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JB2QHS34}},
  note         = {Machine review of arXiv:2501.16461}
}
read the original abstract

AI systems are increasingly tasked to complete responsibilities with decreasing oversight. This delegation requires users to accept certain risks, typically mitigated by perceived or actual alignment of values between humans and AI, leading to confidence that the system will act as intended. However, stochastic behavior by an AI system threatens to undermine alignment and potential trust. In this work, we take a philosophical perspective to the tension and potential conflict between stochasticity and trustworthiness. We demonstrate how stochasticity complicates traditional methods of establishing trust and evaluate two extant approaches to managing it: (1) eliminating user-facing stochasticity to create deterministic experiences, and (2) allowing users to independently control tolerances for stochasticity. We argue that both approaches are insufficient, as not all forms of stochasticity affect trustworthiness in the same way or to the same degree. Instead, we introduce a novel definition of stochasticity and propose latent value modeling for both AI systems and users to better assess alignment. This work lays a foundational step toward understanding how and when stochasticity impacts trustworthiness, enabling more precise trust calibration in complex AI systems, and underscoring the importance of sociotechnical analyses to effectively address these challenges.

Figures

Figures reproduced from arXiv: 2501.16461 by the authors.

Figure 1
Figure 1. Causal diagrams from User and LLM Perspectives. Yellow nodes are observed elements, red are user-determined latent states [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗
Figure 2
Figure 2. Causal diagram of User and LLM Perspectives Combined. Yellow nodes are observed elements, red are user-determined latent [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

101 extracted references · 72 canonical work pages

  1. [1]

    Improving IoT technology adoption through improving consumer trust

    Areej AlHogail. “Improving IoT technology adoption through improving consumer trust”. In: Technologies 6.3 (2018), p. 64

  2. [2]

    Understanding the Role of Social Influence on Consumer Trust in Adopting AI Tools

    Geeta Sandeep Nadella et al. “Understanding the Role of Social Influence on Consumer Trust in Adopting AI Tools”. In: International Journal of Sustainable Development in Computing Science 5.2 (2023), pp. 1–18

  3. [3]

    Trust: A requirement for cloud technology adoption

    Akinwale O Akinwunmi, Emmanuel A Olajubu, and G Adesola Aderounmu. “Trust: A requirement for cloud technology adoption”. In: International Journal of Advanced Computer Science and Applications 6.8 (2015), pp. 112–118

  4. [4]

    Trust in and adoption of online recommendation agents

    Izak Benbasat and Weiquan Wang. “Trust in and adoption of online recommendation agents”. In: Journal of the association for information systems 6.3 (2005), p. 4

  5. [5]

    Trust and the market for technology

    Paul H Jensen, Alfons Palangkaraya, and Elizabeth Webster. “Trust and the market for technology”. In:Research Policy 44.2 (2015), pp. 340–356

  6. [6]

    Vulnerability, Trust and AI

    Ahmer Arif and Os Keyes. “Vulnerability, Trust and AI”. In: Proceedings of 2022 CHI Workshop on Trust and Reliance in AI-Human Teams. 2022

  7. [7]

    Algorithmic censoring in dynamic learning systems

    Jennifer Chien, Margaret Roberts, and Berk Ustun. “Algorithmic censoring in dynamic learning systems”. In: Proceedings of the 3rd ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization . 2023, pp. 1–20

  8. [8]

    Machine learning and reproducibility impact of random numbers

    David Hill et al. “Machine learning and reproducibility impact of random numbers”. In: (2024)

Show all 101 references
  1. [9]

    Reproducibility in learning

    Russell Impagliazzo et al. “Reproducibility in learning”. In:Proceedings of the 54th annual ACM SIGACT symposium on theory of computing . 2022, pp. 818–831

  2. [10]

    Reproducibility in optimization: Theoretical framework and limits

    Kwangjun Ahn et al. “Reproducibility in optimization: Theoretical framework and limits”. In:Advances in Neural Information Processing Systems 35 (2022), pp. 18022–18033

  3. [11]

    Tesla settles with Apple Engineer’s family who said autopilot caused his fatal crash | CNN business

    David Goldman. Tesla settles with Apple Engineer’s family who said autopilot caused his fatal crash | CNN business . Apr. 2024. url: https://www.cnn.com/2024/04/08/tech/tesla-trial-wrongful-death-walter-huang/index.html

  4. [12]

    Inside the final seconds of a deadly Tesla Autopilot crash

    Trisha Thadani et al. Inside the final seconds of a deadly Tesla Autopilot crash . Oct. 2023. url: https://www. washingtonpost.com/technology/interactive/2023/tesla-autopilot-crash-analysis/

  5. [13]

    Enhancing Human Creativity with Aptly Uncontrollable Generative AI

    Iikka Hauhio. “Enhancing Human Creativity with Aptly Uncontrollable Generative AI”. In: Proceedings of the 15th International Conference on Computational Creativity. Association for Computational Creativity (ACC) . 2024

  6. [14]

    Generation Probabilities Are Not Enough: Uncertainty Highlighting in AI Code Completions

    Helena Vasconcelos et al. “Generation Probabilities Are Not Enough: Uncertainty Highlighting in AI Code Completions”. In: ACM Transactions on Computer-Human Interaction (2024)

  7. [15]

    Quality-diversity through ai feedback

    Herbie Bradley et al. “Quality-diversity through ai feedback”. In: arXiv preprint arXiv:2310.13032 (2023)

  8. [16]

    Generative ai meets open-ended survey responses: Participant use of ai and homogenization

    Simone Zhang, Janet Xu, and A Alvero. Generative ai meets open-ended survey responses: Participant use of ai and homogenization. 2024

  9. [17]

    Contextual Confidence and Generative AI

    Shrey Jain, Zoë Hitzig, and Pamela Mishkin. “Contextual Confidence and Generative AI”. In: arXiv preprint arXiv:2311.01193 (2023)

  10. [18]

    AI safety on whose terms? 2023

    Seth Lazar and Alondra Nelson. AI safety on whose terms? 2023

  11. [19]

    Fairness and abstraction in sociotechnical systems

    Andrew D Selbst et al. “Fairness and abstraction in sociotechnical systems”. In: Proceedings of the conference on fairness, accountability, and transparency. 2019, pp. 59–68

  12. [20]

    Trust—The importance of trustfulness versus trustworthiness

    Jan Tullberg. “Trust—The importance of trustfulness versus trustworthiness”. In: The journal of socio-economics 37.5 (2008), pp. 2059–2071. Manuscript submitted to ACM Trustworthiness in Stochastic Systems: Towards Opening the Black Box 17

  13. [21]

    Trust as a commodity

    Partha Dasgupta. “Trust as a commodity”. In:Trust: Making and breaking cooperative relations 4 (2000), pp. 49–72

  14. [22]

    Trustworthiness and trust: influences and implications

    Harjit Sekhon et al. “Trustworthiness and trust: influences and implications”. In: Journal of marketing manage- ment 30.3-4 (2014), pp. 409–430

  15. [23]

    Not so different after all: A cross-discipline view of trust

    Denise M. Rousseau et al. “Not so different after all: A cross-discipline view of trust”. In: The Academy of Management Review 23 (1998), pp. 393–404

  16. [24]

    Trust and antitrust

    Annette C. Baier. “Trust and antitrust”. In: Ethics 96 (1986), pp. 231–260

  17. [25]

    Building trust by signaling trustworthiness in service retail

    Husni Kharouf, Donald J. Lund, and Harjit Sekhon. “Building trust by signaling trustworthiness in service retail”. In: Journal of Services Marketing 28.5 (2014), pp. 361–373

  18. [26]

    Should Users Trust Advanced AI Assistants? Justified Trust As a Function of Competence and Alignment

    Arianna Manzini et al. “Should Users Trust Advanced AI Assistants? Justified Trust As a Function of Competence and Alignment”. In: The 2024 ACM Conference on Fairness, Accountability, and Transparency. 2024, pp. 1174–1186

  19. [27]

    Trust and human-machine teaming: A qualitative study

    Joseph B Lyons et al. “Trust and human-machine teaming: A qualitative study”. In: Artificial intelligence for the internet of everything. Elsevier, 2019, pp. 101–116

  20. [28]

    Institutional quality and generalized trust: A nonrecursive causal model

    Blaine G Robbins. “Institutional quality and generalized trust: A nonrecursive causal model”. In:Social indicators research 107 (2012), pp. 235–258

  21. [29]

    Does human-robot trust need reciprocity?

    Joshua Zonca and Alessandra Sciutti. “Does human-robot trust need reciprocity?” In:arXiv preprint arXiv:2110.09359 (2021)

  22. [30]

    The concept of trustworthiness: A cross-cultural comparison between Japanese and US business people

    Masami Nishishiba and L David Ritchie. “The concept of trustworthiness: A cross-cultural comparison between Japanese and US business people”. In: (2000)

  23. [31]

    Contextual influences on trust and trustworthiness: An etic perspective

    Catherine T Kwantes and Suzanne McMurphy. “Contextual influences on trust and trustworthiness: An etic perspective”. In: Trust and trustworthiness across cultures: Implications for societies and workplaces (2021), pp. 1– 15

  24. [32]

    For your local eyes only: Culture-specific face typicality influences perceptions of trustwor- thiness

    Carmel Sofer et al. “For your local eyes only: Culture-specific face typicality influences perceptions of trustwor- thiness”. In: Perception 46.8 (2017), pp. 914–928

  25. [33]

    Does trust beget trustworthiness? Trust and trustworthiness in two games and two cultures: A research note

    Toko Kiyonari et al. “Does trust beget trustworthiness? Trust and trustworthiness in two games and two cultures: A research note”. In: Social psychology quarterly 69.3 (2006), pp. 270–283

  26. [34]

    In AI we trust: ethics, artificial intelligence, and reliability

    Mark Ryan. “In AI we trust: ethics, artificial intelligence, and reliability”. In: Science and Engineering Ethics 26.5 (2020), pp. 2749–2767

  27. [35]

    Human values, free will, and the conscious mind

    George Edgin Pugh. “Human values, free will, and the conscious mind”. In: Zygon: Journal of Religion and Science 11.1 (1976)

  28. [36]

    Moral molecules: Morality as a combinatorial system

    Oliver Scott Curry et al. “Moral molecules: Morality as a combinatorial system”. In: Review of Philosophy and Psychology 13.4 (2022), pp. 1039–1058

  29. [37]

    Stability and change in work values: A meta-analysis of longitudinal studies

    Jing Jin and James Rounds. “Stability and change in work values: A meta-analysis of longitudinal studies”. In: Journal of Vocational Behavior 80.2 (2012), pp. 326–339

  30. [38]

    The structure of intraindividual value change

    Anat Bardi et al. “The structure of intraindividual value change.” In: Journal of personality and social psychology 97.5 (2009), p. 913

  31. [39]

    Competing values in organizations: Contextual influences and structural consequences

    Victoria Buenger et al. “Competing values in organizations: Contextual influences and structural consequences”. In: Organization Science 7.5 (1996), pp. 557–576

  32. [40]

    Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries

    Shalom H Schwartz. “Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries”. In: Advances in experimental social psychology/Academic Press (1992)

  33. [41]

    Modeling and analysis of stochastic systems

    Vidyadhar G Kulkarni. Modeling and analysis of stochastic systems . Chapman and Hall/CRC, 2016

  34. [42]

    Overreliance on AI literature review

    Samir Passi and Mihaela Vorvoreanu. “Overreliance on AI literature review”. In: Microsoft Research (2022). Manuscript submitted to ACM 18 Jennifer Chien and David Danks

  35. [43]

    “I’m Not Sure, But

    Sunnie SY Kim et al. ““I’m Not Sure, But... ”: Examining the Impact of Large Language Models’ Uncertainty Ex- pression on User Reliance and Trust”. In:The 2024 ACM Conference on Fairness, Accountability, and Transparency. 2024, pp. 822–835

  36. [44]

    You Can Only Verify When You Know the Answer: Feature-Based Explanations Reduce Overreliance on AI for Easy Decisions, but Not for Hard Ones

    Zelun Tony Zhang et al. “You Can Only Verify When You Know the Answer: Feature-Based Explanations Reduce Overreliance on AI for Easy Decisions, but Not for Hard Ones”. In: Proceedings of Mensch und Computer

  37. [45]

    Explaining the Unexplainable: The Impact of Misleading Explanations on Trust in Unreliable Predictions for Hardly Assessable Tasks

    Mersedeh Sadeghi et al. “Explaining the Unexplainable: The Impact of Misleading Explanations on Trust in Unreliable Predictions for Hardly Assessable Tasks”. In:Proceedings of the 32nd ACM Conference on User Modeling, Adaptation and Personalization. 2024, pp. 36–46

  38. [46]

    Explanations can reduce overreliance on ai systems during decision-making

    Helena Vasconcelos et al. “Explanations can reduce overreliance on ai systems during decision-making”. In: Proceedings of the ACM on Human-Computer Interaction 7.CSCW1 (2023), pp. 1–38

  39. [47]

    The impact, advancements and applications of generative AI

    Balagopal Ramdurai and Prasanna Adhithya. “The impact, advancements and applications of generative AI”. In: International Journal of Computer Science and Engineering 10.6 (2023), pp. 1–8

  40. [48]

    Designing socio-technical systems

    Johannes M Bauer and Paulien M Herder. “Designing socio-technical systems”. In: Philosophy of technology and engineering sciences. Elsevier, 2009, pp. 601–630

  41. [49]

    Insights into the nature of technology diffusion and implementation: the perspective of sociotechnical alignment

    Alfonso H Molina. “Insights into the nature of technology diffusion and implementation: the perspective of sociotechnical alignment”. In: Technovation 17.11-12 (1997), pp. 601–626

  42. [50]

    Preference inconsistency in multidisciplinary design decision making

    Erin F MacDonald, Richard Gonzalez, and Panos Y Papalambros. “Preference inconsistency in multidisciplinary design decision making”. In: (2009)

  43. [51]

    ‘Interpretability’and ‘alignment’are fool’s errands: a proof that controlling misaligned large language models is the best anyone can hope for

    Marcus Arvan. “‘Interpretability’and ‘alignment’are fool’s errands: a proof that controlling misaligned large language models is the best anyone can hope for”. In: AI & SOCIETY (2024), pp. 1–16

  44. [52]

    Why google took down Gemini’s AI image generator and the drama around IT - The Washington Post

    Gerrit De Vynck and Nitasha Tiku. Why google took down Gemini’s AI image generator and the drama around IT - The Washington Post. Feb. 2024. url: https://www.washingtonpost.com/technology/2024/02/22/google-gemini- ai-image-generation-pause/

  45. [53]

    The moral limits of predictive practices: The case of credit-based insurance scores

    Barbara Kiviat. “The moral limits of predictive practices: The case of credit-based insurance scores”. In:American Sociological Review 84.6 (2019), pp. 1134–1158

  46. [54]

    The Algorithmic Link Between Auto Insurance Pricing and Unfair Discrimination

    Dorothy L Andrews. “The Algorithmic Link Between Auto Insurance Pricing and Unfair Discrimination”. PhD thesis. Fielding Graduate University, 2023

  47. [55]

    A review of the role of artificial intelligence in healthcare

    Ahmed Al Kuwaiti et al. “A review of the role of artificial intelligence in healthcare”. In: Journal of personalized medicine 13.6 (2023), p. 951

  48. [56]

    Trust and medical AI: the challenges we face and the expertise needed to overcome them

    Thomas P Quinn et al. “Trust and medical AI: the challenges we face and the expertise needed to overcome them”. In: Journal of the American Medical Informatics Association 28.4 (2021), pp. 890–894

  49. [57]

    Platform governance and the “infodemic

    Eugenia Siapera. “Platform governance and the “infodemic””. In: Javnost-The Public 29.2 (2022), pp. 197–214

  50. [58]

    Argumentation strategies in lobbying: the discursive struggle over proposals to regulate Big Tech

    Scott Davidson and Irina Lock. “Argumentation strategies in lobbying: the discursive struggle over proposals to regulate Big Tech”. In: Journal of Communication Management (2024)

  51. [59]

    The tech lobby: Tracing the contours of new media elite lobbying power

    Pawel Popiel. “The tech lobby: Tracing the contours of new media elite lobbying power”. In: Communication Culture & Critique 11.4 (2018), pp. 566–585

  52. [60]

    Big Techs and Global Financial Regulation: Intersection, Challenges, and Solutions

    Steve Kourabas and Cheng-Yun CY Tsang. “Big Techs and Global Financial Regulation: Intersection, Challenges, and Solutions”. In: Challenges, and Solutions (August 30, 2023) 40 (2023)

  53. [61]

    Harnessing The Science Of Persuasion. pdf

    Robert B Cialdini. “Harnessing The Science Of Persuasion. pdf”. In: (2001). Manuscript submitted to ACM Trustworthiness in Stochastic Systems: Towards Opening the Black Box 19

  54. [62]

    One vs. Many: Comprehending Accurate Information from Multiple Erroneous and In- consistent AI Generations

    Yoonjoo Lee et al. “One vs. Many: Comprehending Accurate Information from Multiple Erroneous and In- consistent AI Generations”. In: The 2024 ACM Conference on Fairness, Accountability, and Transparency . 2024, pp. 2518–2531

  55. [63]

    Customer satisfaction of recommender system: Examining accuracy and diversity in several types of recommendation approaches

    Jaekyeong Kim, Ilyoung Choi, and Qinglong Li. “Customer satisfaction of recommender system: Examining accuracy and diversity in several types of recommendation approaches”. In: Sustainability 13.11 (2021), p. 6165

  56. [64]

    How AI Ideas Affect the Creativity, Diversity, and Evolution of Human Ideas: Evidence From a Large, Dynamic Experiment

    Joshua Ashkinaze et al. “How AI Ideas Affect the Creativity, Diversity, and Evolution of Human Ideas: Evidence From a Large, Dynamic Experiment”. In: arXiv preprint arXiv:2401.13481 (2024)

  57. [65]

    The Role of Generative AI in Human Creative Processes: Experimental Evidence

    Feng Zhu and Wenbo Zou. “The Role of Generative AI in Human Creative Processes: Experimental Evidence”. In: A vailable at SSRN 4676053(2023)

  58. [66]

    When ChatGPT is gone: Creativity reverts and homogeneity persists

    Qinghan Liu et al. “When ChatGPT is gone: Creativity reverts and homogeneity persists”. In: arXiv preprint arXiv:2401.06816 (2024)

  59. [67]

    How AI can distort human beliefs

    Celeste Kidd and Abeba Birhane. “How AI can distort human beliefs”. In: Science 380.6651 (2023), pp. 1222–1223

  60. [68]

    The placebo effect of artificial intelligence in human–computer interaction

    Thomas Kosch et al. “The placebo effect of artificial intelligence in human–computer interaction”. In: ACM Transactions on Computer-Human Interaction 29.6 (2023), pp. 1–32

  61. [69]

    Human Learning about AI Performance

    Bnaya Dreyfuss and Raphael Raux. “Human Learning about AI Performance”. In:arXiv preprint arXiv:2406.05408 (2024)

  62. [70]

    Beyond Behaviorist Representational Harms: A Plan for Measurement and Mitigation

    Jennifer Chien and David Danks. “Beyond Behaviorist Representational Harms: A Plan for Measurement and Mitigation”. In: The 2024 ACM Conference on Fairness, Accountability, and Transparency . 2024, pp. 933–946

  63. [71]

    AI hallucination: towards a comprehensive classification of distorted information in artificial intelligence-generated content

    Yujie Sun et al. “AI hallucination: towards a comprehensive classification of distorted information in artificial intelligence-generated content”. In: Humanities and Social Sciences Communications 11.1 (2024), pp. 1–14

  64. [72]

    The Paradox of Collective Certainty in Science

    Eamon Duede and James Evans. “The Paradox of Collective Certainty in Science”. In:arXiv preprint arXiv:2406.05809 (2024)

  65. [73]

    When (ish) is my bus? user-centered visualizations of uncertainty in everyday, mobile predictive systems

    Matthew Kay et al. “When (ish) is my bus? user-centered visualizations of uncertainty in everyday, mobile predictive systems”. In: Proceedings of the 2016 chi conference on human factors in computing systems . 2016, pp. 5092–5103

  66. [74]

    Enhancing User Perception of Reliability in Computer Vision: Uncertainty Visualization for Probability Distributions

    Xinyue Wang, Ruoyu Hu, and Chengqi Xue. “Enhancing User Perception of Reliability in Computer Vision: Uncertainty Visualization for Probability Distributions”. In: Symmetry 16.8 (2024), p. 986

  67. [75]

    Why authors don’t visualize uncertainty

    Jessica Hullman. “Why authors don’t visualize uncertainty”. In: IEEE transactions on visualization and computer graphics 26.1 (2019), pp. 130–139

  68. [76]

    Improving privacy settings control in online social networks with a wheel interface

    Tziporah Stern and Nanda Kumar. “Improving privacy settings control in online social networks with a wheel interface”. In: Journal of the Association for Information Science and Technology 65.3 (2014), pp. 524–538

  69. [77]

    Impact of menu complexity upon user behavior and satisfaction in information search

    Svetlana S Bodrunova and Alexandr Yakunin. “Impact of menu complexity upon user behavior and satisfaction in information search”. In: Human Interface and the Management of Information. Information in Applications and Services: 20th International Conference, HIMI 2018, Held as ...

  70. [78]

    A comparison of static, adaptive, and adaptable menus

    Leah Findlater and Joanna McGrenere. “A comparison of static, adaptive, and adaptable menus”. In: Proceedings of the SIGCHI conference on Human factors in computing systems . 2004, pp. 89–96

  71. [79]

    Measurement and fairness

    Abigail Z Jacobs and Hanna Wallach. “Measurement and fairness”. In: Proceedings of the 2021 ACM conference on fairness, accountability, and transparency . 2021, pp. 375–385

  72. [80]

    What Is Fairness

    Jim Dator, D Pratt, and Y Seo. “What Is Fairness”. In: Fairness, Globalization, and Public Institutions 19 (2006). Manuscript submitted to ACM 20 Jennifer Chien and David Danks

  73. [81]

    Training a helpful and harmless assistant with reinforcement learning from human feedback

    Yuntao Bai et al. “Training a helpful and harmless assistant with reinforcement learning from human feedback”. In: arXiv preprint arXiv:2204.05862 (2022)

  74. [82]

    RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs

    Shreyas Chaudhari et al. “RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs”. In: arXiv preprint arXiv:2404.08555 (2024)

  75. [83]

    Strong and weak alignment of large language models with human values

    Mehdi Khamassi, Marceau Nahon, and Raja Chatila. “Strong and weak alignment of large language models with human values”. In: Scientific Reports 14.1 (2024), p. 19399

  76. [84]

    Open problems and fundamental limitations of reinforcement learning from human feedback

    Stephen Casper et al. “Open problems and fundamental limitations of reinforcement learning from human feedback”. In: arXiv preprint arXiv:2307.15217 (2023)

  77. [85]

    LLM is Like a Box of Chocolates: the Non-determinism of ChatGPT in Code Generation

    Shuyin Ouyang et al. “LLM is Like a Box of Chocolates: the Non-determinism of ChatGPT in Code Generation”. In: arXiv preprint arXiv:2308.02828 (2023)

  78. [86]

    The expectation-maximization algorithm

    Todd K Moon. “The expectation-maximization algorithm”. In: IEEE Signal processing magazine 13.6 (1996), pp. 47–60

  79. [87]

    Markov chain Monte Carlo method and its application

    Stephen Brooks. “Markov chain Monte Carlo method and its application”. In: Journal of the royal statistical society: series D (the Statistician) 47.1 (1998), pp. 69–100

  80. [88]

    A comprehensive overview of large language models

    Humza Naveed et al. “A comprehensive overview of large language models”. In: arXiv preprint arXiv:2307.06435 (2023)

  81. [89]

    Values stability and change in adulthood: A 3-year longitudinal study of rank-order stability and mean-level differences

    Taciano L Milfont, Petar Milojev, and Chris G Sibley. “Values stability and change in adulthood: A 3-year longitudinal study of rank-order stability and mean-level differences”. In: Personality and Social Psychology Bulletin 42.5 (2016), pp. 572–588

  82. [90]

    Stability and change of basic personal values in early adulthood: An 8-year longitudinal study

    Michele Vecchione et al. “Stability and change of basic personal values in early adulthood: An 8-year longitudinal study”. In: Journal of Research in Personality 63 (2016), pp. 111–122

  83. [91]

    Rank-order consistency and profile stability of self-and informant-reports of personal values in comparison to personality traits

    Henrik Dobewall and Toivo Aavik. “Rank-order consistency and profile stability of self-and informant-reports of personal values in comparison to personality traits”. In: Journal of Individual Differences (2016)

  84. [92]

    Implications of human value shift and persistence for biodiversity conservation

    Michael J Manfredo, Tara L Teel, and Alia M Dietsch. “Implications of human value shift and persistence for biodiversity conservation”. In: Conservation Biology 30.2 (2016), pp. 287–296

  85. [93]

    Changes in personal values in pandemic times

    Ella Daniel et al. “Changes in personal values in pandemic times”. In:Social Psychological and Personality Science 13.2 (2022), pp. 572–582

  86. [94]

    Patterns of value change during the life span: Some evidence from a functional approach to values

    Valdiney V Gouveia et al. “Patterns of value change during the life span: Some evidence from a functional approach to values”. In: Personality and Social Psychology Bulletin 41.9 (2015), pp. 1276–1290

  87. [95]

    Personal values before and after migration: A longitudinal case study on value change in Ingrian–Finnish migrants

    Jan-Erik Lönnqvist, Inga Jasinskaja-Lahti, and Markku Verkasalo. “Personal values before and after migration: A longitudinal case study on value change in Ingrian–Finnish migrants”. In: Social Psychological and Personality Science 2.6 (2011), pp. 584–591

  88. [96]

    Changes in values and well-being amidst the COVID-19 pandemic in Poland

    Agnieszka Bojanowska et al. “Changes in values and well-being amidst the COVID-19 pandemic in Poland”. In: PloS one 16.9 (2021), e0255491

  89. [97]

    Brief report: Early adolescents’ value development at war time

    Ella Daniel et al. “Brief report: Early adolescents’ value development at war time”. In: Journal of adolescence 36.4 (2013), pp. 651–655

  90. [98]

    Changes in young Europeans’ values during the global financial crisis

    Florencia M Sortheix et al. “Changes in young Europeans’ values during the global financial crisis”. In: Social Psychological and Personality Science 10.1 (2019), pp. 15–25

  91. [99]

    Values following a major terrorist incident: Finnish adolescent and student values before and after September 11, 2001

    Markku Verkasalo, Robin Goodwin, and Irina Bezmenova. “Values following a major terrorist incident: Finnish adolescent and student values before and after September 11, 2001”. In: Journal of Applied Social Psychology 36.1 (2006), pp. 144–160. Manuscript submitted to ACM Trustw...

  92. [100]

    It Takes Both Trust and Lack of Mistrust: The Workings of Cooperation and Relational Signaling in Contractual Relationships

    Siegwart Lindenberg. “It Takes Both Trust and Lack of Mistrust: The Workings of Cooperation and Relational Signaling in Contractual Relationships.” In: Journal of Management & Governance 4 (2000)

  93. [101]

    Artificial general intelligence: concept, state of the art, and future prospects

    Ben Goertzel. “Artificial general intelligence: concept, state of the art, and future prospects”. In: Journal of Artificial General Intelligence 5.1 (2014), p. 1. Manuscript submitted to ACM 22 Jennifer Chien and David Danks A STOCHASTICITY AND TRUSTWORTHINESS IN INTERPERSONAL...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.