Pith. sign in

REVIEW 3 major objections 5 minor 66 references

How trust networks shape students' opinions about the proficiency of artificially intelligent assistants

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Trust networks, not just test scores, decide whether students perceive an AI assistant's proficiency correctly, and one stubbornly wrong student can prevent a class from ever converging on the truth.

desk verdict Solid reapplication of the authors' opinion-dynamics framework to AI proficiency perceptions, but the key 'indefinite disruption' claim needs longer runs or a stability argument. read the letter →

arxiv 2506.19655 v1 pith:ULCIEW63 submitted 2025-06-24 physics.soc-ph

Yutong Bu , Andrew Melatos , Robin Evans This is my paper · ORCID
classification physics.soc-ph
keywords opiniondynamicsperceivedAIproficiencytrustnetworksmulti-agentsimulationBayesianlearningpartisansasymptoticineducation
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

The paper argues that what students believe about an AI assistant's proficiency is shaped as much by who they trust as by the scores they actually receive. Using Monte Carlo simulations of a probabilistic opinion-dynamics model, it claims that in an all-allies classroom everyone eventually infers the true proficiency $ heta_{\rm AI}$, that students who do not use AI converge faster than those who do, and that adding a single stubbornly wrong student destroys convergence entirely, leaving everyone's beliefs vacillating indefinitely between the true value and the partisan's value. In low-trust classrooms the opposite holds: everyone settles down, but only a minority lands on the truth. Mixed trust networks produce turbulent nonconvergence and intermittency that depend delicately on network wiring. These results matter because perceived proficiency drives AI adoption, and the model predicts that equal access to the tool is not enough to guarantee accurate shared judgments.

What carries the argument

Each student's belief is a probability density function $x_i(t,\theta)$ over a discretized proficiency axis $\theta\in\{0,0.05,\dots,1\}$. At every time step the belief is updated in two steps: first, AI-users apply Bayes's theorem to their assessment score, whose Gaussian likelihood is centered on the true $\theta_{\rm AI}$; second, all students share their PDFs and mix them linearly through the trust matrix $A_{ij}\in\{+1,0,-1\}$, which pulls a student's belief toward allies and pushes it away from opponents via the average deviation $\Delta x'_i(t+1/2,\theta)$, with learning rate $\mu=0.25$ and renormalization. A partisan is a node whose PDF is fixed at a single value $\theta_{\rm p}$. The results come from the competition between the multiplicative Bayesian evidence step and the additive trust-mixing step: a fixed wrong node acts as a permanent source of attraction that direct evidence cannot erase, while opponents-only networks stabilize because repulsion cancels information flow rather than amplifying it.

What would settle it

Take a real ten-person study group where everyone shares probabilistic beliefs openly and one planted participant never changes a wrong estimate of the AI tool's proficiency; if the group's averaged belief converges to the AI's true proficiency rather than vacillating indefinitely, the paper's main nonconvergence claim is contradicted.

Watch

Extended reading notes

Core claim

The central claim is that perceived proficiency of an AI tool is a socially emergent quantity, not merely an individual inference from direct evidence. Under the model's update rule, an allies-only network converges to the correct $ heta_{\rm AI}=0.8$ for all students, with AI-avoiders reaching asymptotic learning faster than AI-users because they rely on the network-averaged belief, which fluctuates less than individual scores. However, one partisan student holding a fixed wrong belief $\theta_{\rm p}\neq\theta_{\rm AI}$ makes every other student's belief bimodal and nonconvergent, even though AI-users observe repeated direct evidence. In opponents-only networks all students reach asymptotic learning, yet only a minority (about 16 percent in the simulations) correctly infer $ heta_{\rm AI}$, and AI-users hold a small but real advantage over AI-avoiders. Mixed networks can leave some students never settling, some alternating between stable and unstable phases, and the long-term outcome depends sensitively on the pattern of trust relationships; these qualitative patterns persist in networks up to about 1000 students.

Load-bearing premise

The predictions all rest on the assumption that students openly share their complete probability distributions with fixed, symmetric, ternary trust relations; if real students communicate only summaries, hide beliefs, or change whom they trust, the nonconvergence and learning-time results may not appear.

Editorial extensions

If this is right

  • In an allies-only classroom with no partisans, all students eventually learn the AI tool's true proficiency, and AI-avoiders learn it faster than AI-users.
  • A single stubbornly wrong student or teacher is sufficient to stop the whole network from reaching a settled opinion, even when abundant direct evidence about the true proficiency is available.
  • In low-trust environments, students reach stable but often wrong conclusions, with only a minority identifying the true proficiency and AI-users enjoying a slight edge over AI-avoiders.
  • Mixed trust networks can produce individuals who never converge, or who alternate between stable and unstable phases, so the same AI tool can be judged accurately by some students and inaccurately by others within one class.
  • The main qualitative results persist in larger networks up to about 1000 students, and a partisan teacher allied with everyone is more influential in large sparse networks than in small dense ones.

Reading between the lines

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

  • If these simulations transfer to real classrooms, publishing independent, unbiased proficiency benchmarks is a necessary but not sufficient remedy: the model implies that one respected wrong voice can outweigh abundant evidence.
  • The AI-avoiders-learn-faster result suggests a possible second-mover advantage in AI adoption, where students who watch peers' outcomes rather than experimenting directly may form more stable judgments; the paper does not test this adoption implication directly.
  • A natural, untested extension is to relax full transparency: if students share only summaries of their beliefs rather than complete PDFs, or if trust relationships evolve over time, the partisan disruption could weaken or strengthen; that is an inference beyond the paper's assumptions.
  • The same machinery could apply to any educational resource whose quality is inferred socially, such as textbooks, but only when students consult it in an oracular way; the paper itself flags this boundary condition.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper presents a multi-agent Monte Carlo model of opinion dynamics in which students form probabilistic beliefs about the proficiency of an AI assistant. Beliefs update via a two-step rule: a Bayesian update from the student's own assessment scores (if the student is an AI-user) and a non-Bayesian linear peer-pressure step that moves beliefs toward the PDFs of trusted allies and away from those of mistrusted opponents (Eqs. 1-6). The authors simulate allies-only, opponents-only, and mixed networks of size n=10, and also larger networks (n=100 to 1000), with and without partisan students or a partisan teacher. The main reported findings are: in allies-only networks all students eventually infer the AI proficiency theta_AI, with AI-avoiders reaching asymptotic learning faster than AI-users; in opponents-only networks a minority infer theta_AI correctly, with a small AI-user advantage; in mixed networks turbulent nonconvergence and intermittency occur; and a single partisan with an incorrect belief theta_p != theta_AI causes persistent vacillation in allies-only networks, described in the abstract as 'indefinite'. The paper concludes with tentative educational policy implications.

Significance. If the central claims hold, the paper offers a striking and counterintuitive result: in a high-trust classroom, one stubbornly misinformed student can prevent the whole cohort from converging on the true proficiency of an AI tool, despite repeated direct evidence. The finding extends the authors' prior media-bias model to an educational context and adds a new, emergent result (AI-avoiders learning faster than AI-users). The model is clearly specified, the simulation methodology is transparent, and the authors are honest about many limitations, explicitly deferring various generalizations. However, the most important claim ('indefinitely' vacillation) rests on finite-horizon simulations only, and at least one numerical result is internally contradictory; these issues currently prevent the paper from establishing its headline conclusions at the level of rigor expected for a physics journal.

major comments (3)
  1. [Abstract and Section 4.1] The claim that a single partisan with theta_p != theta_AI makes students' beliefs 'vacillate indefinitely' between theta_p and theta_AI is not established by the simulations, which are truncated at T=10^4. The convergence criterion in Eq. (8) with tau_max=99 can be satisfied for long windows by a slowly drifting bimodal PDF, so the observed nonconvergence within the run does not rule out convergence on a longer timescale. The paper should either provide a longer-run stability analysis (e.g., T=10^6, or a spectral/Lyapunov analysis of the update map in Eqs. (1)-(6)) or soften the wording to 'over the simulation horizon' in the abstract and conclusions.
  2. [Section 5.1 and Figure 6] There is a direct contradiction between the text and the caption of Figure 6 regarding AI-avoiders in large allies-only networks. The text states that AI-avoiders achieve asymptotic learning with ta in the range 49.3 to 66.7 for n=200, while the caption states that AI-avoiders are not plotted because all AI-avoiders fail to achieve asymptotic learning by t=T=10^4. This inconsistency undermines the claimed persistence of the AI-avoider learning-time advantage in larger networks and must be resolved before the large-network results can be accepted.
  3. [Section 2.2, Eq. (5)] The choice mu=0.25 is stated 'without loss of generality', but no sensitivity analysis over mu is provided. Since mu controls the strength of peer pressure relative to direct observation, the qualitative outcomes (notably turbulent nonconvergence and the relative learning times of AI-users and AI-avoiders) may depend on mu. Please either demonstrate insensitivity across the allowed range 0<mu<=0.5, or cite a previous systematic study that establishes the robustness of these specific results to mu.
minor comments (5)
  1. [Section 3.1] The prior beliefs are only described as 'randomized', without specifying the distribution (e.g., uniform over the simplex, Dirichlet, etc.). This hampers reproducibility; please specify the prior generation procedure.
  2. [Section 3.1, text near Fig. 1] The sentence 'For example, for n=1, the histogram peaks at ta...' should read 'for k=1' (since n=10 throughout that experiment); as written it is confusing.
  3. [Section 4.1, paragraph on the second suite of tests] The statement that 'we do not observe that the modal ta decreases monotonically with k' is unsupported, because the reported partisan simulations use only k=1; no variation of k is presented in that subsection.
  4. [Section 5.3, bottom right panel] The observation that for theta_p=theta_AI in a Barabasi-Albert network with n=1000 no agent achieves asymptotic learning within T=10^4 appears to contradict the small-network result where a correct partisan accelerates learning; the paper defers longer runs but gives no explanation for the discrepancy, leaving the reader to wonder whether the timescale for larger networks simply exceeds T or whether the qualitative behavior differs.
  5. [Section 3.2, top right panel description] The violin plots in Fig. 2 show distributions of differences in asymptotic learning time; it would be helpful to state explicitly how many simulations had undefined differences (e.g., when no student reaches the right or wrong conclusion), and how those cases are handled in the plot.

Circularity Check

1 steps flagged · score 4.0 of 10

Central partisan-disruption result is imported from the authors' own Ref [31]; the AI-avoider speed result is emergent, but the asymptotic disruption claim rests on a self-citation chain.

  1. self citation load bearing [Section 4.1 (Partisan student), with model choices in Section 2.2 and Eq. (8); also Abstract.]
    "This phenomenon agrees with the counterintuitive finding in Ref. [31], that a single partisan destabilizes an allies-only network regardless of its size, stopping everyone from achieving asymptotic learning. ... We copy the successful implementation strategy for partisans described in Section 2.2 of Ref. [31]."

    The paper's headline claim that a single partisan makes beliefs 'vacillate indefinitely between θp and θAI' is not re-derived in the present paper; it is explicitly taken from the authors' own prior work, Ref [31]. The partisan mechanism is not independently specified here: the implementation is copied from Ref [31], and the stationarity criterion in Eq. (8) is also described as 'copying the arbitrary choice made in Ref. [29, 30, 31]'. The in-paper Monte Carlo runs stop at T = 10^4, so the asymptotic 'indefinitely' component of the central claim is a self-citation chain plus finite-time extrapolation rather than an in-paper derivation. This makes the most load-bearing conclusion of the paper depend on the authors' own earlier result.

full rationale

The model in Eqs. (1)-(6) is the authors' own Bayesian-PDF plus signed-trust update from Refs [29, 30, 31], with the hidden quantity relabeled from media bias to AI proficiency. Reusing one's own model is not circular by itself, and no parameters are fitted to the target outcomes: θAI = 0.8, σAI = 0.24, and μ = 0.25 are set a priori, and no real-world data are predicted from fitted values. The new 'AI-avoiders reach asymptotic learning faster than AI-users' result is an emergent averaging consequence of the update rule, not an input. The genuine circularity is concentrated in the partisan-disruption headline, which is explicitly carried over from the authors' own Ref [31], and the asymptotic 'indefinitely' wording exceeds what the T = 10^4 simulations in this paper alone can establish. The finite-time nonconvergence is a correctness and evidence concern rather than circularity by itself, but the self-citation is load-bearing for the paper's most prominent claim. The educational application and the AI-user/AI-avoider comparisons provide independent content, so a moderately elevated score of 4 is appropriate rather than a higher score that would imply the whole derivation reduces to its inputs.

Assumptions & free parameters 9 free parameters · 7 assumptions · 0 invented entities

The model rests on a set of explicit domain assumptions about score generation, belief sharing, and network statics. No new physical or conceptual entities are introduced beyond the opinion-dynamics framework already present in the authors' prior work. The free parameters are chosen by hand, not fitted to data, and several are not subjected to sensitivity analysis.

free parameters (9)
  • learning rate μ = 0.25
    Set in Section 2.2 'without loss of generality' but not systematically varied; affects convergence properties of Eq. (5).
  • true AI proficiency θ_AI = 0.8
    Injected value used in all simulations; results for other values are only briefly explored in partisan cases.
  • AI score noise σ_AI = 0.24
    Standard deviation of AI scores in Eq. (2); controls likelihood width and affects learning times.
  • consensus tolerance ε = 0.01
    User-selected threshold in Eqs. (7)-(8) for defining consensus and asymptotic learning; changes the classification of outcomes.
  • time window τ_max = 99
    Time window in the asymptotic-learning criterion Eq. (8), copied from prior work; arbitrary and influences which agents are considered converged.
  • discretization grid points = 21
    θ discretized into 21 equally spaced values (Section 3); coarse grid makes convergence criteria easier to satisfy.
  • simulation horizon T = 10^4
    Maximum simulation length; some nonconvergence claims (e.g., AI-users in n=200 allies-only) are relative to this finite horizon.
  • network sizes n = 10, 200, 1000
    Class sizes studied; the paper claims qualitative results depend weakly on n, but quantitative outcomes may vary.
  • Barabási-Albert attachment m = 3
    Parameter for scale-free networks; only one value tested, affecting connectivity and convergence speeds.
assumptions (7)
  • domain assumption AI-users' scores are drawn from a truncated Gaussian with mean θ_AI and standard deviation σ_AI (Eq. (2)).
    Assumed score-generating process used in simulations; if real AI score distributions are non-Gaussian or biased, learning dynamics differ.
  • domain assumption AI-avoiders' scores contain no information about θ_AI; their likelihood is unity (Eq. (3)).
    This makes AI-avoiders entirely dependent on peer information; any leakage of AI proficiency into their own scores would alter the dynamics.
  • domain assumption Beliefs update via Bayes' rule followed by the linear signed peer-pressure rule with max(0, ...) rectification and renormalization (Eqs. (1)-(6)).
    The specific functional form of social influence is assumed; other update rules (e.g., bounded confidence) could change results.
  • domain assumption Trust relationships are static, symmetric, take values +1/0/-1, and are known to all (Eq. (4)).
    Real trust networks are dynamic, asymmetric, and weighted; the model's results depend on this simplification.
  • domain assumption Students share complete belief PDFs openly with all network neighbors (Section 2.2).
    This full transparency is load-bearing: the peer-pressure mechanism in Eq. (5) relies on exact PDF sharing.
  • domain assumption Partisans hold a fixed single belief θp (delta-function PDF) and are completely immune to social influence (Section 4).
    The disruption result depends on this extreme obduracy; more flexible partisans might not cause permanent vacillation.
  • domain assumption Students never change their AI-usage strategy; AI-users and AI-avoiders are fixed for the entire simulation (Section 2.1).
    The findings do not account for strategic switching in response to performance or social pressure, which could alter convergence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of How trust networks shape students' opinions about the proficiency of artificially intelligent assistants." pith.science (2026). https://pith.science/paper/ULCIEW63

@misc{pith2026250619655,
  author       = {Pith},
  title        = {Pith review of: How trust networks shape students' opinions about the proficiency of artificially intelligent assistants},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ULCIEW63}},
  note         = {Machine review of arXiv:2506.19655}
}
abstract

The rising use of educational tools controlled by artificial intelligence (AI) has provoked a debate about their proficiency. While intrinsic proficiency, especially in tasks such as grading, has been measured and studied extensively, perceived proficiency remains underexplored. Here it is shown through Monte Carlo multi-agent simulations that trust networks among students influence their perceptions of the proficiency of an AI tool. A probabilistic opinion dynamics model is constructed, in which every student's perceptions are described by a probability density function (PDF), which is updated at every time step through independent, personal observations and peer pressure shaped by trust relationships. It is found that students infer correctly the AI tool's proficiency $\theta_{\rm AI}$ in allies-only networks (i.e.\ high trust networks). AI-avoiders reach asymptotic learning faster than AI-users, and the asymptotic learning time for AI-users decreases as their number increases. However, asymptotic learning is disrupted even by a single partisan, who is stubbornly incorrect in their belief $\theta_{\rm p} \neq \theta_{\rm AI}$, making other students' beliefs vacillate indefinitely between $\theta_{\rm p}$ and $\theta_{\rm AI}$. In opponents-only (low trust) networks, all students reach asymptotic learning, but only a minority infer $\theta_{\rm AI}$ correctly. AI-users have a small advantage over AI-avoiders in reaching the right conclusion. In mixed networks, students may exhibit turbulent nonconvergence and intermittency, or achieve asymptotic learning, depending on the relationships between partisans and AI-users. The educational implications of the results are discussed briefly in the context of designing robust usage policies for AI tools, with an emphasis on the unintended and inequitable consequences which arise sometimes from counterintuitive network effects.

Figures

Figures reproduced from arXiv: 2506.19655 by the authors.

Figure 1
Figure 1. Correctly inferring 𝜃AI in an allies-only network with 𝑛 = 10. (Top.) Mean belief ⟨𝜃⟩ versus time 𝑡 for a representative AI-user (blue curve) and AI-avoider (pink curve). Both curves asymptote to the correct value of 𝜃AI. (Bottom left.) Semi-violin plot of the distribution of asymptotic learning time 𝑡a as a function of the number of AI-users in a complete network. Histograms of 𝑡a are separated into AI-users (𝑏𝑖(𝑡)… view at source ↗
Figure 2
Figure 2. Incorrectly inferring 𝜃AI in an opponents-only network with 𝑛 = 10. (Top left.) Mean belief ⟨𝜃⟩ versus time 𝑡 for AI-users (blue curve) and AI-avoiders (pink curve). Two of the AI-users and none of the AI-avoiders infer 𝜃AI correctly. (Top right.) Violin plot of the distribution of the difference in asymptotic learning time ⟨𝑡 right a ⟩−⟨𝑡 wrong a ⟩ between students who do and do not correctly infer 𝜃AI, averaged ov… view at source ↗
Figure 3
Figure 3. Turbulent nonconvergence and intermittency in a mixed network, for a typical simulation of a complete network with [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Disruption by a partisan student in an allies-only complete network (top panels) and a mixed complete network (bottom panel) with [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Disruption by a partisan teacher in a mixed complete teacher-student network with [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]
Figure 6
Figure 6. Figure 6: Correctly inferring 𝜃AI in a larger allies-only network with 𝑛 = 200, to be compared with the smaller network with 𝑛 = 10 in [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]
Figure 7
Figure 7. Figure 7: Incorrectly inferring 𝜃AI in a large opponents-only complete network with 𝑛 = 200. (Left.) Same as the top left panel in [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: Disruption by a partisan student in two larger ( [PITH_FULL_IMAGE:figures/full_fig_p019_8.png]
Figure 9
Figure 9. Figure 9: Opinion formation in larger classes: influence of a teacher in a mixed Barabási-Albert network with 100 students, zero or one partisan [PITH_FULL_IMAGE:figures/full_fig_p020_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 40 canonical work pages

  1. [31]

    Y. Bu, A. Melatos, Discerning media bias within a network of political allies and opponents: Disruption by partisans, Physica A: Statistical Mechanics and its Applications 624 (2023) 128958. doi:10.1016/j.physa.2023.128958. URL https://www.sciencedirect.com/science/article/pii/S0378437123005137 2, 3, 5, 6, 7, 8, 9, 11, 12, 22, 23

    work page arXiv 2023

  2. [1]

    Crompton, D

    H. Crompton, D. Burke, Artificial intelligence in higher education: the state of the field, International Journal of Educational Technology in Higher Education 20 (1) (2023) 22. doi:10.1186/s41239-023-00392-8 . URL https://doi.org/10.1186/s41239-023-00392-8 1

    doi:10.1186/s41239-023-00392-8 2023

  3. [2]

    van den Berg, E

    G. van den Berg, E. du Plessis, ChatGPT and Generative AI: Possibilities for Its Contribution to Lesson Planning, Critical Thinking and Openness in Teacher Education, Education Sciences 13 (10) (2023) 998, number: 10 Publisher: Multidisciplinary Digital Publishing Institute. doi:10.3390/educsci13100998. URL https://www.mdpi.com/2227-7102/13/10/998 1

    doi:10.3390/educsci13100998 2023

  4. [3]

    R. Liu, C. Zenke, C. Liu, A. Holmes, P. Thornton, D. J. Malan, Teaching CS50 with AI: Leveraging Generative Artificial Intelligence in Computer Science Education, in: Proceedings of the 55th ACM Technical Symposium on Computer Science Education V. 1, SIGCSE 2024, Association for Computing Machinery, New York, NY, USA, 2024, pp. 750–756. doi:10.1145/362625...

    arXiv 2024

  5. [4]

    S. M. Rutner, R. A. Scott, Use of Artificial Intelligence to Grade Student Discussion Boards: An Exploratory Study, Information Systems Education Journal 20 (4) (2022) 4–18, publisher: Information Systems and Computing Academic Professionals ERIC Number: EJ1358299. URL https://eric.ed.gov/?id=EJ1358299 1

    work page 2022

  6. [5]

    C. W. Okonkwo, A. Ade-Ibijola, Chatbots applications in education: A systematic review, Computers and Education: Artificial Intelligence 2 (2021) 100033. doi:10.1016/j.caeai.2021.100033. URL https://www.sciencedirect.com/science/article/pii/S2666920X21000278 1

    arXiv 2021

  7. [6]

    Labadze, M

    L. Labadze, M. Grigolia, L. Machaidze, Role of AI chatbots in education: systematic literature review, International Journal of Educational Technology in Higher Education 20 (1) (2023) 56. doi:10.1186/s41239-023-00426-1 . URL https://doi.org/10.1186/s41239-023-00426-1 1

    doi:10.1186/s41239-023-00426-1 2023

  8. [7]

    Vučković, S

    D. Vučković, S. Peković, M. Blečić, R. Ðoković, Attitudes towards cheating behavior during assessing students performance: student and teacher perspectives, International Journal for Educational Integrity 16 (1) (2020) 13. doi:10.1007/s40979-020-00065-3 . URL https://doi.org/10.1007/s40979-020-00065-3 1

    work page doi:10.1007/s40979-020-00065-3 2020

Show all 66 references
  1. [8]

    Hwang, H

    G.-J. Hwang, H. Xie, B. W. Wah, D. Gašević, Vision, challenges, roles and research issues of Artificial Intelligence in Education, Computers and Education: Artificial Intelligence 1 (2020) 100001. doi:10.1016/j.caeai.2020.100001. URL https://www.sciencedirect.com/science/artic...

    2020

  2. [9]

    Bharadiya, Artificial Intelligence and the Future of Web 3.0: Opportunities and Challenges Ahead, American Journal of Computer Science and Technology 6 (2023) 74–79

    J. Bharadiya, Artificial Intelligence and the Future of Web 3.0: Opportunities and Challenges Ahead, American Journal of Computer Science and Technology 6 (2023) 74–79. doi:10.11648/j.ajcst.20230602.14. 2

    2023 doi

  3. [10]

    H. O. Khogali, S. Mekid, The blended future of automation and AI: Examining some long-term societal and ethical impact features, Tech- nology in Society 73 (2023) 102232. doi:10.1016/j.techsoc.2023.102232. URL https://www.sciencedirect.com/science/article/pii/S0160791X23000374 2

    2023

  4. [11]

    M. J. Reiss, The use of AI in education: Practicalities and ethical considerations, London Review of Education 19 (1) (2021). doi:10. 14324/LRE.19.1.05. URL https://journals.uclpress.co.uk/lre/article/id/1267/ 2

    2021

  5. [12]

    S. Jeon, Y. S. Chang, S. J. Jo, T. Madukuand, Y. E. Kim, Speed of Catch-Up and Convergence of the Artificial Intelligence Divide: AI Investment, Robotic, Start-Ups, and Patents, Journal of Global Information Technology Management 27 (1) (2024) 63–85, publisher: Routledge _epri...

    2024

  6. [13]

    Sharma, T

    H. Sharma, T. Soetan, T. Farinloye, E. Mogaji, M. D. F. Noite, AI Adoption in Universities in Emerging Economies: Prospects, Challenges and Recommendations, in: E. Mogaji, V. Jain, F. Maringe, R. E. Hinson (Eds.), Re-imagining Educational Futures in Developing Countries: Lesso...

    2022 doi

  7. [14]

    Casal-Otero, A

    L. Casal-Otero, A. Catala, C. Fernández-Morante, M. Taboada, B. Cebreiro, S. Barro, AI literacy in K-12: a systematic literature review, International Journal of STEM Education 10 (1) (2023) 29. doi:10.1186/s40594-023-00418-7 . URL https://doi.org/10.1186/s40594-023-00418-7 2

    2023 doi

  8. [15]

    E. Ferrara, Fairness and Bias in Artificial Intelligence: A Brief Survey of Sources, Impacts, and Mitigation Strategies, Sci 6 (1) (2024) 3, number: 1 Publisher: Multidisciplinary Digital Publishing Institute. doi:10.3390/sci6010003. URL https://www.mdpi.com/2413-4155/6/1/3 2

    2024 doi

  9. [16]

    T. E. Commission, Ethical guidelines on the use of artificial intelligence (AI) and data in teaching and learning for Educators. 2

  10. [17]

    C. K. Y. Chan, A comprehensive AI policy education framework for university teaching and learning, International Journal of Educational Technology in Higher Education 20 (1) (2023) 38. doi:10.1186/s41239-023-00408-3 . URL https://doi.org/10.1186/s41239-023-00408-3 2

    2023 doi

  11. [18]

    Zhang, S

    M. Zhang, S. Baral, N. Heffernan, A. Lan, Automatic Short Math Answer Grading via In-context Meta-learningPublisher: [object Object] Version Number: 3 (2022). doi:10.48550/ARXIV.2205.15219. URL https://arxiv.org/abs/2205.15219 2

    2022 doi arXiv

  12. [19]

    G. Kortemeyer, Toward AI grading of student problem solutions in introductory physics: A feasibility study, Physical Review Physics Edu- cation Research 19 (2) (2023) 020163, publisher: American Physical Society.doi:10.1103/PhysRevPhysEducRes.19.020163. URL https://link.aps.or...

    2023 doi

  13. [20]

    Vijaya Shetty, K

    S. Vijaya Shetty, K. R. Guruvyas, P. P. Patil, J. J. Acharya, Essay Scoring Systems Using AI and Feature Extraction: A Review, in: V. Bindhu, J. M. R. S. Tavares, K.-L. Du (Eds.), Proceedings of Third International Conference on Communication, Computing and Electronics Systems...

    2022 doi

  14. [21]

    E. Hall, M. Seyam, D. Dunlap, Identifying Usability Challenges in AI-Based Essay Grading Tools, in: N. Wang, G. Rebolledo-Mendez, V. Dimitrova, N. Matsuda, O. C. Santos (Eds.), Artificial Intelligence in Education. Posters and Late Breaking Results, Workshops and 25 Tutorials,...

    2023 doi

  15. [22]

    J. H. Choi, K. E. Hickman, A. Monahan, D. Schwarcz, ChatGPT Goes to Law School (Jan. 2023). doi:10.2139/ssrn.4335905. URL https://papers.ssrn.com/abstract=4335905 2

    2023 doi

  16. [23]

    Meaney, R

    C. Meaney, R. S. Huang, K. J. Q. Lu, A. W. Fischer, F.-H. Leung, K. Kulasegaram, K. Tzanetos, A. Punnett, Comparing the Performance of ChatGPT and GPT-4 versus a Cohort of Medical Students on an Official University of Toronto Undergraduate Medical Education Progress Test, page...

    2023 doi

  17. [24]

    Z. Ji, N. Lee, R. Frieske, T. Yu, D. Su, Y. Xu, E. Ishii, Y. J. Bang, A. Madotto, P. Fung, Survey of Hallucination in Natural Language Generation, ACM Computing Surveys 55 (12) (2023) 248:1–248:38. doi:10.1145/3571730. URL https://dl.acm.org/doi/10.1145/3571730 2

    2023 doi

  18. [25]

    Why Should I Trust You?

    M. T. Ribeiro, S. Singh, C. Guestrin, "Why Should I Trust You?": Explaining the Predictions of Any Classifier, in: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ACM, San Francisco California USA, 2016, pp. 1135–

    2016

  19. [26]

    Achiam, S

    OpenAI, J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat, R. Avila, I. Babuschkin, S. Balaji, V. Balcom, P. Baltescu, H. Bao, M. Bavarian, J. Belgum, I. Bello, J. Berdine, G. Bernadett-Shapiro, C. Berner, L....

    2024 arXiv

  20. [27]

    Y. Wang, C. Liu, Y.-F. Tu, Factors Affecting the Adoption of AI-Based Applications in Higher Education: An Analysis of Teachers Perspectives Using Structural Equation Modeling, Educational Technology & Society 24 (3) (2021) 116–129, publisher: International Forum of Educationa...

    2021

  21. [28]

    Ye, Opinion Dynamics and the Evolution of Social Power in Social Networks, Springer, 2019, google-Books-ID: PNaIDwAAQBAJ

    M. Ye, Opinion Dynamics and the Evolution of Social Power in Social Networks, Springer, 2019, google-Books-ID: PNaIDwAAQBAJ. 2

    2019

  22. [29]

    N. K. Y. Low, A. Melatos, Discerning media bias within a network of political allies and opponents: The idealized example of a biased coin, Physica A: Statistical Mechanics and its Applications 590 (2022) 126722. doi:10.1016/j.physa.2021.126722. URL https://www.sciencedirect.c...

    2022

  23. [30]

    N. Low, A. Melatos, Vacillating about media bias: Changing one’s mind intermittently within a network of political allies and opponents, Physica A: Statistical Mechanics and its Applications 604 (2022). doi:10.1016/j.physa.2022.127829. 2, 3, 5, 6, 9, 22, 23

    2022

  24. [32]

    Deffuant, D

    G. Deffuant, D. Neau, F. Amblard, G. Weisbuch, Mixing Beliefs Among Interacting Agents, Advances in Complex Systems 3 (2000) 87–98. doi:10.1142/S0219525900000078. 2, 3, 5

    2000 doi

  25. [33]

    M. H. Degroot, Reaching a Consensus, Journal of the American Statistical Association 69 (345) (1974) 118–121, publisher: Taylor & Francis _eprint: https://www.tandfonline.com/doi/pdf/10.1080/01621459.1974.10480137. doi:10.1080/01621459.1974.10480137. URL https://www.tandfonlin...

    1974

  26. [34]

    A. Fang, K. Yuan, J. Geng, X. Wei, Opinion Dynamics with Bayesian Learning, Complexity 2020 (2020) e8261392, publisher: Hindawi. doi:10.1155/2020/8261392. URL https://www.hindawi.com/journals/complexity/2020/8261392/ 2, 3, 5

    2020 doi

  27. [35]

    G. Shi, A. Proutiere, M. Johansson, J. S. Baras, K. H. Johansson, The Evolution of Beliefs over Signed Social Networks, Operations Research 64 (3) (2016) 585–604, publisher: INFORMS. doi:10.1287/opre.2015.1448. URL https://pubsonline.informs.org/doi/10.1287/opre.2015.1448 2, 5

    2016

  28. [36]

    A. C. R. Martins, CONTINUOUS OPINIONS AND DISCRETE ACTIONS IN OPINION DYNAMICS PROBLEMS, International Journal of 26 Modern Physics C 19 (04) (2008) 617–624. doi:10.1142/S0129183108012339. URL https://www.worldscientific.com/doi/abs/10.1142/S0129183108012339 2

    2008 doi

  29. [37]

    V. Sood, S. Redner, Voter Model on Heterogeneous Graphs, Physical Review Letters 94 (17) (2005) 178701, publisher: American Physical Society. doi:10.1103/PhysRevLett.94.178701. URL https://link.aps.org/doi/10.1103/PhysRevLett.94.178701 3, 5

    2005 doi

  30. [38]

    Jadbabaie, P

    A. Jadbabaie, P. Molavi, A. Sandroni, A. Tahbaz-Salehi, Non-Bayesian social learning, Games and Economic Behavior 76 (1) (2012) 210–225. doi:10.1016/j.geb.2012.06.001. URL https://www.sciencedirect.com/science/article/pii/S0899825612000851 3, 5

    2012 doi

  31. [39]

    A. Fang, L. Wang, X. Wei, Social learning with multiple true states, Physica A: Statistical Mechanics and its Applications 521 (2019) 375–386. doi:10.1016/j.physa.2019.01.089. URL https://www.sciencedirect.com/science/article/pii/S037843711930086X 3, 6

    2019 doi

  32. [40]

    Kaufmann, P

    T. Kaufmann, P. Weng, V. Bengs, E. Hüllermeier, A Survey of Reinforcement Learning from Human Feedback, arXiv:2312.14925 [cs] (Dec. 2023). URL http://arxiv.org/abs/2312.14925 4

    2023

  33. [41]

    Huang, S

    J. Huang, S. S. Gu, L. Hou, Y. Wu, X. Wang, H. Yu, J. Han, Large Language Models Can Self-Improve, arXiv:2210.11610 [cs] (Oct. 2022). doi:10.48550/arXiv.2210.11610. URL http://arxiv.org/abs/2210.11610 4

    2022 doi arXiv

  34. [42]

    T. B. Brown, B. Mann, N. Ryder, M. Subbiah, J. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, A. Askell, S. Agarwal, A. Herbert- Voss, G. Krueger, T. Henighan, R. Child, A. Ramesh, D. M. Ziegler, J. Wu, C. Winter, C. Hesse, M. Chen, E. Sigler, M. Litwin, S. Gray, B....

    2020 arXiv

  35. [43]

    Jackson, Social and Economic Networks, Princeton University Press, 2008

    M. Jackson, Social and Economic Networks, Princeton University Press, 2008. 4

    2008

  36. [44]

    Betz, Natural-Language Multi-Agent Simulations of Argumentative Opinion Dynamics, Journal of Artificial Societies and Social Simu- lation 25 (1) (2022) 2

    G. Betz, Natural-Language Multi-Agent Simulations of Argumentative Opinion Dynamics, Journal of Artificial Societies and Social Simu- lation 25 (1) (2022) 2. 5

    2022

  37. [45]

    M. G. Koeslag-Kreunen, M. R. Van der Klink, P. Van den Bossche, W. H. Gijselaers, Leadership for team learning: the case of university teacher teams, Higher Education 75 (2) (2018) 191–207. doi:10.1007/s10734-017-0126-0 . URL https://doi.org/10.1007/s10734-017-0126-0 5

    2018 doi

  38. [46]

    Y. Li, S. Ma, Y. Zhang, R. Huang, Kinshuk, An improved mix framework for opinion leader identification in online learning communities, Knowledge-Based Systems 43 (2013) 43–51. doi:10.1016/j.knosys.2013.01.005. URL https://www.sciencedirect.com/science/article/pii/S0950705113000099 5

    2013 doi

  39. [47]

    J. Zhou, C. Huang, Q. Dai, The effects of conformity-driven teaching ability on opinion consensus, Europhysics Letters 123 (3) (2018) 30004, publisher: EDP Sciences, IOP Publishing and Società Italiana di Fisica. doi:10.1209/0295-5075/123/30004. URL https://dx.doi.org/10.1209/...

    2018 doi

  40. [48]

    Vaz Martins, M

    T. Vaz Martins, M. Pineda, R. Toral, Mass media and repulsive interactions in continuous-opinion dynamics, EPL (Europhysics Letters) 91 (4) (2010) 48003. doi:10.1209/0295-5075/91/48003. URL https://iopscience.iop.org/article/10.1209/0295-5075/91/48003 5

    2010 doi

  41. [49]

    X. Chen, P. Tsaparas, J. Lijffijt, T. De Bie, Opinion Dynamics with Backfire Effect and Biased Assimilation, Tech. Rep. arXiv:1903.11535, arXiv, arXiv:1903.11535 [cs] type: article (Mar. 2019). doi:10.48550/arXiv.1903.11535. URL http://arxiv.org/abs/1903.11535 5

    1903 doi arXiv

  42. [50]

    G. He, J. Liu, H. Hu, J.-A. Fang, Discrete-time signed bounded confidence model for opinion dynamics, Neurocomputing 425 (2021) 53–61. doi:10.1016/j.neucom.2019.12.061. URL https://www.sciencedirect.com/science/article/pii/S0925231219317709 5

    2021 doi

  43. [51]

    J. Tee, D. P. Taylor, A Quantized Representation of Probability in the Brain, IEEE Transactions on Molecular, Biological and Multi-Scale Communications 5 (1) (2019) 19–29, conference Name: IEEE Transactions on Molecular, Biological and Multi-Scale Communications.doi: 10.1109/T...

    2019

  44. [52]

    A. Fan, M. Lewis, Y. Dauphin, Hierarchical Neural Story Generation, in: I. Gurevych, Y. Miyao (Eds.), Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Association for Computational Linguistics, Melbourne, Australi...

    2018 doi

  45. [53]

    Holtzman, J

    A. Holtzman, J. Buys, L. Du, M. Forbes, Y. Choi, The Curious Case of Neural Text Degeneration, 2020. URL https://iclr.cc/virtual_2020/poster_rygGQyrFvH.html 6

    2020

  46. [54]

    Radford, J

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever, Language Models are Unsupervised Multitask Learners (2019). 6

    2019

  47. [55]

    C. Yang, X. Wang, Y. Lu, H. Liu, Q. V. Le, D. Zhou, X. Chen, Large Language Models as Optimizers, arXiv:2309.03409 [cs] (Apr. 2024). URL http://arxiv.org/abs/2309.03409 6

    2024 arXiv

  48. [56]

    K. Chen, A. Shao, J. Burapacheep, Y. Li, Conversational AI and equity through assessing GPT-3’s communication with diverse so- cial groups on contentious topics, Scientific Reports 14 (1) (2024) 1561, publisher: Nature Publishing Group. doi:10.1038/ s41598-024-51969-w . URL ht...

    2024

  49. [57]

    Lalitha, T

    A. Lalitha, T. Javidi, A. D. Sarwate, Social Learning and Distributed Hypothesis Testing, IEEE Transactions on Information Theory 64 (9) (2018) 6161–6179, conference Name: IEEE Transactions on Information Theory. doi:10.1109/TIT.2018.2837050. URL https://ieeexplore.ieee.org/do...

    2018

  50. [58]

    D. M. Rothschild, J. Wolfers, Forecasting Elections: Voter Intentions Versus Expectations (Jul. 2011). doi:10.2139/ssrn.1884644. URL https://papers.ssrn.com/abstract=1884644 7

    2011 doi

  51. [59]

    Heider, Attitudes and Cognitive Organization, The Journal of Psychology 21 (1) (1946) 107–112, publisher: Routledge _eprint: https://doi.org/10.1080/00223980.1946.9917275

    F. Heider, Attitudes and Cognitive Organization, The Journal of Psychology 21 (1) (1946) 107–112, publisher: Routledge _eprint: https://doi.org/10.1080/00223980.1946.9917275. doi:10.1080/00223980.1946.9917275. 27 URL https://doi.org/10.1080/00223980.1946.9917275 9, 13

    1946

  52. [60]

    Cartwright, F

    D. Cartwright, F. Harary, Structural balance: a generalization of Heider’s theory, Psychological Review 63 (1956) 277–293, place: US Pub- lisher: American Psychological Association. doi:10.1037/h0046049. 9, 13

    1956 doi

  53. [61]

    Mobilia, A

    M. Mobilia, A. Petersen, S. Redner, On the role of zealotry in the voter model, Journal of Statistical Mechanics: Theory and Experiment 2007 (08) (2007) P08029–P08029, publisher: IOP Publishing. doi:10.1088/1742-5468/2007/08/P08029. URL https://doi.org/10.1088/1742-5468/2007/0...

    2007 doi

  54. [62]

    Mobilia, Does a Single Zealot Affect an Infinite Group of Voters?, Physical Review Letters 91 (2) (2003) 028701, publisher: American Physical Society

    M. Mobilia, Does a Single Zealot Affect an Infinite Group of Voters?, Physical Review Letters 91 (2) (2003) 028701, publisher: American Physical Society. doi:10.1103/PhysRevLett.91.028701. URL https://link.aps.org/doi/10.1103/PhysRevLett.91.028701 11, 12

    2003 doi

  55. [63]

    Mobilia, I

    M. Mobilia, I. T. Georgiev, Voting and catalytic processes with inhomogeneities, Physical Review E 71 (4) (2005) 046102, publisher: American Physical Society. doi:10.1103/PhysRevE.71.046102. URL https://link.aps.org/doi/10.1103/PhysRevE.71.046102 12

    2005 doi

  56. [64]

    K. R. Ankney, G. J. Hidding, Fast-Follower Advantages and Network Externalities in I.T.-Driven Markets, 2005. URL https://www.semanticscholar.org/paper/Fast-Follower-Advantages-and-Network-Externalities-Ankney-Hidding/ c9ac6f5fc8a8f420527ea8efb272b9f551cd0fa2 21

    2005

  57. [65]

    Kopel, C

    M. Kopel, C. Löffler, Commitment, first-mover-, and second-mover advantage, Journal of Economics 94 (2) (2008) 143–166. doi:10. 1007/s00712-008-0004-4 . URL https://doi.org/10.1007/s00712-008-0004-4 21 28 No Partisan Partisan Student Partisan Teacher Trust relationships 𝜃p ≠𝜃A...

    2008 doi

  58. [1144]

    URL https://dl.acm.org/doi/10.1145/2939672.2939778 2

    doi:10.1145/2939672.2939778. URL https://dl.acm.org/doi/10.1145/2939672.2939778 2

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.