Pith. sign in

REVIEW 3 major objections 7 minor 1 cited by

Emotionally-Aware Agents for Dispute Resolution

T0 review · 3 major / 7 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Emotion intensities generated by GPT-4o from dispute text explain more than 40% of the variance in how participants feel about the process and their relationship, far exceeding prior negotiation-based emotion recognition.

desk verdict LLM emotion labels explain a lot of variance in dispute outcomes—but the headline R² is in-sample, prompt-tuned on the test subset, and not clearly distinct from a general negativity factor. read the letter →

arxiv 2509.04465 v1 pith:MEWX4PMY submitted 2025-08-28 cs.CL cs.AI

Sushrita Rakshit , James Hale , Kushal Chawla , Jeanne M. Brett , Jonathan Gratch This is my paper · ORCID
classification cs.CLcs.AI
keywords disputeresolutionemotionrecognitionlargelanguagemodelsaffectivecomputingsubjectivevalueangerescalationcompassionnegotiation
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

Disputes evoke stronger emotions than negotiations, and anger can spiral into retaliation rather than compromise. The paper asks whether automatically recognized emotional expressions from text can predict how people experience a dispute's outcome. Using 2,025 buyer-seller disputes from the KODIS corpus, it shows that prompting GPT-4o to annotate each utterance with soft intensities over anger, joy, fear, sadness, surprise, compassion, and neutral yields labels that, averaged over a dialogue, explain in some cases more than 40% of the variance in the Subjective Value Inventory's four subscales, with buyers' feelings about process and relationship approaching half the variance. This is far above the roughly 5% explained by recognized anger in prior negotiation research, and the same pattern holds across several LLMs that also match human annotators better than the T5 baseline. The labels also trace two dynamics: seller reciprocation of buyer anger tracks impasse, while early seller compassion and buyer reciprocation track resolution.

What carries the argument

The machinery is LLM-based utterance-level emotion intensity annotation followed by linear regression. The paper prompts GPT-4o with dialogue history, a seven-label scheme replacing 'love' with 'compassion' and adding 'neutral', soft labels summing to one, and a small set of in-context examples to score each turn; these intensity vectors are averaged over a dialogue and used as predictors of each SVI subscale. The R-squared of those regressions is the paper's measure of explanatory power. For dynamics, the same annotations are averaged by role and turn to trace anger and compassion trajectories in resolved versus impasse dialogues.

What would settle it

Give the same regression pipeline a single generic negativity or adversarial-tone score per dialogue, such as a lexical hostility measure or a human rating of how negative the conversation was, alongside or instead of the seven emotion labels. If the generic score alone recovers most of the R-squared, or if adding it collapses the emotion coefficients, the specific emotion-reading claim is not supported. Alternatively, collect human soft-label emotion annotations on a large random sample of dialogues and run the same regressions: if human labels explain far less than GPT-4o's labels, the high

Watch

Extended reading notes

Core claim

The central claim is that automatically recognized emotional expression is a measurable and substantial correlate of dispute outcomes. Concretely, the paper reports that when GPT-4o is prompted to produce per-utterance soft emotion-intensity vectors, including dialogue context, a label set of anger, joy, fear, sadness, surprise, compassion, and neutral, and a few in-context examples, the dialogue-level averages predict a disputant's subjective outcome: linear regressions on these labels explain, in some cases, over 40% of the variance in SVI subscales, with buyers' feelings about process and relationship approaching half the variance. The paper further claims that this explanatory power is n

Load-bearing premise

The claim stands or falls on whether GPT-4o's soft emotion labels measure the emotions actually expressed, rather than an overall adversarial tone or generic negativity that happens to correlate with bad outcomes.

Editorial extensions

If this is right

  • Emotion labels alone, without dialogue content, can flag high-risk disputes early: dialogues that end in impasse show anger reciprocation within the first few turns.
  • Mediation agents could use such labels to detect escalation in real time and intervene to defuse anger or encourage compassion.
  • Label design matters: adding a neutral category and swapping love for compassion improved predictive performance, so downstream tasks should shape emotion-taxonomy choices in affective computing.
  • LLM-based emotion recognition outperforms the fine-tuned T5 baseline both in matching human annotations and in explaining subjective outcomes across four LLMs, making it a ready replacement in conflict research pipelines.
  • Disputes should be treated as a distinct genre from negotiations in emotion research: impasse rates and emotion-outcome correlations differ sharply.

Reading between the lines

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

  • A control regression that adds a generic negativity or adversarial-tone score alongside the seven emotion labels would show whether the specific emotion labels carry distinct information; the paper does not include such a control.
  • The causal reading of the emotion-outcome link would require an intervention experiment, such as randomly assigning an AI mediator to encourage compassion in one arm; the correlational design alone cannot rule out that emotions merely mirror an underlying dispute quality.
  • The claim may transfer to other dispute settings such as workplace, family, or legal mediation, but the corpus is a single online role-play scenario in English, so cross-setting generalization is untested.
  • An early-warning system that reads only the first few turns is a concrete design implication: the reported trajectories separate impasse from resolution almost immediately after the first round.
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

3 major / 7 minor

Summary. The paper analyzes 2,025 buyer–seller dispute dialogues from the KODIS corpus, using GPT-4o (and other LLMs) to produce soft emotion-intensity labels for each utterance. Compared with T5-Twitter, the LLM labels correlate better with self-reported frustration and with third-party human annotations on a small sample. The authors then use multiple linear regression to predict the four Subjective Value Inventory (SVI) subscales from dialogue-level mean emotion intensities, reporting R2 values up to roughly 0.4, and they examine turn-level trajectories of anger and compassion for disputes ending in impasse versus resolution. The central claims are that automatically recognized emotions explain substantially more variance in dispute outcomes than in prior negotiation research and that the trajectories reveal an anger-escalation spiral and a compassion pathway to resolution.

Significance. If the central estimate survives proper evaluation, the result is substantial: prior negotiation studies reported roughly 5% variance explained by recognized emotion, whereas this paper reports up to ~40% in disputes, and it applies in a domain—dispute resolution—that is understudied in the affective-computing literature. The paper also demonstrates a methodological advantage of context-aware LLM annotation over a fine-tuned T5 baseline, with machine-checkable comparisons across multiple LLMs and human annotations. These strengths make the contribution potentially valuable for emotion-aware agent design and for social-science theories of conflict escalation. However, the quantitative claims currently rest on in-sample R2 values, a prompt-selection procedure that uses the evaluation subset, and a construct-validity argument that has not ruled out a single negativity/positivity factor.

major comments (3)
  1. [§V-B, Table III, Fig. 3] The headline R2 values are computed in-sample, and the prompt configuration is selected on the same evaluation subset. The text says 'We perform these tests on a 20% subset (N = 406) of the corpus' and Table III then reports mean R2 for the ablated configurations, with the best GPT4o configuration used in Fig. 3. Since each model has 7 emotion-intensity predictors (per side) and there is no held-out split, cross-validation, or significance testing, the reported R2—and especially the improvement over T5—can reflect overfitting and selection. Because the abstract and Discussion rely on the 'over 40% of the variance' claim, please report out-of-sample R2 (e.g., 5-fold CV, or train on the remaining 80% and test on the 20%), separate prompt selection from evaluation, and provide confidence intervals or at least standard errors.
  2. [§V-A, Table II; §VI-B, Fig. 5] The validation evidence does not establish that the LLM emotion labels isolate distinct emotional constructs rather than a single valence/negativity factor. The correlation pattern in Table II—anger +0.54, fear +0.36, joy −0.36, compassion −0.18 with self-reported frustration—is exactly what a unidimensional negativity factor would predict. The SVI regressions in Fig. 3 may therefore be driven by overall adversarial tone rather than by specific emotions. To support the specific anger-spiral and compassion-pathway claims, please test incremental validity: do anger and compassion add significant variance to SVI after controlling for an aggregate negative-tone index (e.g., mean intensity of negative labels or a lexical sentiment score)? Additionally, the human-annotation benchmark uses only 100 utterances with no inter-annotator agreement or per-emotion confidence intervals in Fig. 5; repor
  3. [§V-C, Fig. 4a–4b] The escalation-spiral and compassion-pathway conclusions are based on raw mean trajectories with no inferential test. The text states that sellers 'reciprocate the buyer's anger' in impasse dialogues and 'stick to the script' in resolved dialogues, but no statistical comparison of the trajectories (e.g., an outcome × turn interaction in a mixed-effects model, or per-timepoint tests with multiple-comparison correction) is reported. Given that the labels are themselves generated by GPT-4o, the trajectory differences could reflect annotation artifacts. Please add formal comparisons of the trajectory shapes and report uncertainty around the means.
minor comments (7)
  1. [§III-C] Clarify whether the 19% impasse rate is dyad-level or participant-level; the Discussion says '19% of participants failed to reach an agreement,' which is a different statement.
  2. [§IV-B, Table III] The checkmark columns in Table III are hard to parse. Define what each column corresponds to (IC learning, compassion label, dialogue history, neutral label) and describe the in-context examples: who hand-annotated them, how many, and whether they are from the same corpus.
  3. [§VI-B, Fig. 5] Report the number of annotators per utterance, the total number of annotations (the text says N=336, presumably annotators), and the inter-annotator agreement. Also clarify that 'average human' is the mean over annotations, and add error bars to the correlation coefficients.
  4. [Abstract and §VII] The abstract says the paper investigates 'subjective and objective outcomes,' but the regression analysis only predicts subjective SVI scores; objective resolution (impasse vs. resolution) appears only in descriptive trajectory plots. Please align the wording with what is actually estimated.
  5. [Fig. 1] The dialogue snippet contains 'traud'—if this is meant to be 'fraud,' correct it in the figure; if intentional (as a non-word), mark it as such.
  6. [Author affiliations] Typo: 'University of Southern Calirofnia' should be 'University of Southern California.'
  7. [§VIII] The sentence 'we do not find much evidence of interactions between those demographic factors' is not supported by any reported analysis. Either include the analysis or remove the claim.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline R² is an in-sample selected maximum: the prompt is tuned on the same 20% subset used to report the 'over 40% of variance' result.

  1. fitted input called prediction [Section V-B, Table III and Fig. 3]
    "To justify our prompting approach, we first contrast T5-Twitter and the various prompting techniques — ablating or modifying various aspects of the prompt — to see which method's annotations better explain the subjective outcomes. We perform these tests on a 20% subset (N = 406) of the corpus. We use multiple linear regression (MLR) to predict subjective feelings for each role, regressing each of the four SVI sub-scales on the average emotion intensities from a given dialogue. ... Fig. 3 compares this highest-performing GPT configuration with T5."

    The 'highest-performing' GPT configuration is selected by comparing R² values on the same 20% subset (N=406) on which Fig. 3's R² values are then reported. The prompt choices (dialogue context, compassion label, neutral label, in-context learning) are hyperparameters selected to maximize the exact outcome-explanation metric that is subsequently presented as the paper's headline result. Thus the 'over 40% of variance' figure is a selected maximum from fits evaluated on the data used for selection, not an independent prediction on held-out data. Calling this 'predict' masks that the reported explanatory power is statistically forced by the selection procedure; no disjoint validation split is introduced before the headline R² is claimed.

full rationale

The paper does not exhibit definitional circularity: the LLM emotion labels are produced without access to SVI outcomes, and the human-annotation benchmark in Section VI-B is an external ground truth. The self-citations (e.g., Chawla et al.'s T5 baseline and 5% negotiation result, the KODIS corpus) are used as empirical inputs or baselines, not as premises that deduce the paper's conclusions. The main circularity concern is the prompt-selection procedure in Section V-B: the paper tunes the prompting scheme on the very 20% subset from which it then reports the high R² values (Fig. 3 and the 'over 40%' claim). This is a fitted input being called a prediction, and it directly inflates the central quantitative claim. The later benchmarking on N=400 dialogues (Section VI) provides some independent support for the general finding that LLMs outperform T5, and the human-annotation results are independent, so the paper is not entirely circular; however, the headline 'over 40% of variance' is not a clean out-of-sample prediction. Score 6 reflects this partial, load-bearing statistical circularity.

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

The central claim depends on the validity of LLM emotion intensities and on regression fits that are evaluated in-sample. The regression coefficients and prompt selection are free parameters fitted to the evaluation data. No new theoretical entities are introduced; the emotion label set and outcome scales come from prior literature.

free parameters (2)
  • MLR regression coefficients per emotion category and SVI subscale = not reported
    Each SVI subscale is regressed on average emotion intensities; coefficients are fit to the data, and R2 is computed on the same data (in-sample).
  • Prompt configuration (context, in-context examples, label set) = best variant: GPT4o with dialogue context, compassion label, neutral label, in-context learning
    Selected based on R2 on the same 20% subset (N=406) used for the reported comparison, so the configuration is fitted to the evaluation data.
assumptions (5)
  • domain assumption LLM soft-label vectors sum to one and act as comparable intensity measures across utterances and dialogues.
    Section IV-B: GPT is prompted to allocate weights summing to one over seven emotion labels; the analysis treats these as emotion intensities without calibration beyond the small validation sets.
  • domain assumption Averaging per-utterance emotion intensities over a dialogue produces a meaningful dialogue-level emotion summary for regression.
    Section V-B: MLR uses mean emotion intensity per dialogue as predictors; this discards turn-level dynamics except in the separate trajectory plots.
  • domain assumption The 2-item frustration subscale from participants' self-reports is a valid external criterion for validating emotion annotations.
    Section V-A: Correlates emotion averages with self-reported frustration; self-report is retrospective and may reflect post-outcome feelings.
  • domain assumption The simulated online buyer-seller dispute scenario elicits emotions and processes comparable to real-world disputes.
    The KODIS corpus is a paid role-play task, not a real dispute. Section IX acknowledges limited purview and uncertainty about generalization.
  • domain assumption The SVI scale measures subjective value in four dimensions as claimed, and these subscales are appropriate outcome measures.
    Section III-B adopts SVI from prior work without further validation in the dispute context.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Emotionally-Aware Agents for Dispute Resolution." pith.science (2026). https://pith.science/paper/MEWX4PMY

@misc{pith2026250904465,
  author       = {Pith},
  title        = {Pith review of: Emotionally-Aware Agents for Dispute Resolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MEWX4PMY}},
  note         = {Machine review of arXiv:2509.04465}
}
read the original abstract

In conflict, people use emotional expressions to shape their counterparts' thoughts, feelings, and actions. This paper explores whether automatic text emotion recognition offers insight into this influence in the context of dispute resolution. Prior work has shown the promise of such methods in negotiations; however, disputes evoke stronger emotions and different social processes. We use a large corpus of buyer-seller dispute dialogues to investigate how emotional expressions shape subjective and objective outcomes. We further demonstrate that large-language models yield considerably greater explanatory power than previous methods for emotion intensity annotation and better match the decisions of human annotators. Findings support existing theoretical models for how emotional expressions contribute to conflict escalation and resolution and suggest that agent-based systems could be useful in managing disputes by recognizing and potentially mitigating emotional escalation.

Figures

Figures reproduced from arXiv: 2509.04465 by the authors.

Figure 1
Figure 1. A KODIS dialogue snippet illustrating how an initial focus on interests escalates into appeals to moral norms and ultimately threats of retaliation, resulting in an impasse. using large-language models (LLMs) to simulate emotion￾laden interactions [50]. Rehearsal recognizes when statements likely escalate the conflict and provides personalized feedback on more cooperative dialogue moves — results suggest the system … view at source ↗
Figure 3
Figure 3. R2 values for various regressions measuring the effi￾cacy of the emotion labels to predict SVI scores. . or modifying various aspects of the prompt — to see which method’s annotations better explain the subjective outcomes. We perform these tests on a 20% subset (N = 406) of the corpus. We use multiple linear regression (MLR) to predict subjective feelings for each role, regressing each of the four SVI sub-scales on… view at source ↗
Figure 2
Figure 2. Comparison of mean intensity of recognized emotions [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Average expressed anger and compassion across dialogue turns broken out by role and objective outcome. cases explaining almost half of the variance. Still, if we want agents to intervene and steer the conversation away from an impasse, they must understand how expressi…
Figure 5
Figure 5. Figure 5: Emotion intensity correlations with human annotations [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Fit of emotion labels predicting subjective outcome — [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Dialogue Summarization with Emotion Dynamics Using Topic- and Participant-Centric Decomposition

    cs.CL 2026-07 conditional novelty 5.0 of 10

    Combining topic- and participant-level summaries with explicit emotion labels preserves dialogue emotion trajectories better than either view alone, measured by new trajectory metrics.

Reference graph

Works this paper leans on

68 extracted references · 61 canonical work pages · cited by 1 Pith paper

  1. [1]

    Parkinson, A

    B. Parkinson, A. H. Fischer, and A. S. Manstead, Emotion in social relations: Cultural, group, and interpersonal processes . Psychology press, 2005

    work page 2005

  2. [2]

    How to do things with emotional expressions: The theory of affective pragmatics,

    A. Scarantino, “How to do things with emotional expressions: The theory of affective pragmatics,” Psychological Inquiry , vol. 28, no. 2-3, pp. 165–185, 2017

    work page 2017

  3. [3]

    Managing emotions during team problem solving: Emotional intelligence and conflict resolution,

    P. J. Jordan and A. C. Troth, “Managing emotions during team problem solving: Emotional intelligence and conflict resolution,” in Emotion and performance. CRC Press, 2021, pp. 195–218

    work page 2021

  4. [4]

    Social functions of machine emotional expressions,

    C. M. de Melo, J. Gratch, S. Marsella, and C. Pelachaud, “Social functions of machine emotional expressions,” Proceedings of the IEEE , vol. 111, no. 10, pp. 1382–1397, 2023

    work page 2023

  5. [5]

    Affective grounding in human-robot interaction,

    M. F. Jung, “Affective grounding in human-robot interaction,” in Pro- ceedings of the 2017 ACM/IEEE International Conference on Human- Robot Interaction, 2017, pp. 263–273

    work page 2017

  6. [6]

    Using virtual patients to teach empathy: a randomized controlled study to enhance medical students’ empathic communication,

    A. Foster, N. Chaudhary, T. Kim, J. L. Waller, J. Wong, M. Borish, A. Cordar, B. Lok, and P. F. Buckley, “Using virtual patients to teach empathy: a randomized controlled study to enhance medical students’ empathic communication,” Simulation in Healthcare, vol. 11, no. 3, pp. 181–189, 2016

    work page 2016

  7. [7]

    Survey of emotions in human–robot interactions: Perspectives from robotic psychology on 20 years of research,

    R. Stock-Homburg, “Survey of emotions in human–robot interactions: Perspectives from robotic psychology on 20 years of research,” Inter- national Journal of Social Robotics , vol. 14, no. 2, pp. 389–411, 2022

    work page 2022

  8. [8]

    Bert and fasttext embeddings for auto- matic detection of toxic speech,

    A. G. d’Sa, I. Illina, and D. Fohr, “Bert and fasttext embeddings for auto- matic detection of toxic speech,” in 2020 International Multi-Conference on:“Organization of Knowledge and Advanced Technologies”(OCTA) . IEEE, 2020, pp. 1–5

    work page 2020

Show all 68 references
  1. [9]

    The (moral) language of hate,

    B. Kennedy, P. Golazizian, J. Trager, M. Atari, J. Hoover, A. Mostafazadeh Davani, and M. Dehghani, “The (moral) language of hate,” PNAS nexus, vol. 2, no. 7, p. pgad210, 2023

    2023

  2. [10]

    The emergence and transfor- mation of disputes: Naming, blaming, claiming. . . ,

    W. L. Felstiner, R. L. Abel, and A. Sarat, “The emergence and transfor- mation of disputes: Naming, blaming, claiming. . . ,” in Theoretical and Empirical Studies of Rights . Routledge, 2017, pp. 255–306

    2017

  3. [11]

    When will negotiation agents be able to represent us? the challenges and opportunities for autonomous negotiators

    T. Baarslag, M. Kaisers, E. Gerding, C. M. Jonker, and J. Gratch, “When will negotiation agents be able to represent us? the challenges and opportunities for autonomous negotiators.” International Joint Conferences on Artificial Intelligence, 2017

    2017

  4. [12]

    Negotiation decision functions for autonomous agents,

    P. Faratin, C. Sierra, and N. R. Jennings, “Negotiation decision functions for autonomous agents,” Robotics and Autonomous Systems, vol. 24, no. 3-4, pp. 159–182, 1998

    1998

  5. [13]

    Negotiation and cooperation in multi-agent environments,

    S. Kraus, “Negotiation and cooperation in multi-agent environments,” Artificial intelligence, vol. 94, no. 1-2, pp. 79–97, 1997

    1997

  6. [14]

    Negotiating agents,

    C. M. Jonker, K. V . Hindriks, P. Wiggers, and J. Broekens, “Negotiating agents,” AI Magazine, vol. 33, no. 3, pp. 79–79, 2012

    2012

  7. [15]

    Challenges and main results of the automated negotiating agents competition (anac) 2019,

    R. Aydo ˘gan, T. Baarslag, K. Fujita, J. Mell, J. Gratch, D. De Jonge, Y . Mohammad, S. Nakadai, S. Morinaga, H. Osawa et al., “Challenges and main results of the automated negotiating agents competition (anac) 2019,” in Multi-Agent Systems and Agreement Technologies: 17th Eur...

    2019

  8. [16]

    Negotiation as a challenge problem for virtual humans,

    J. Gratch, D. DeVault, G. M. Lucas, and S. Marsella, “Negotiation as a challenge problem for virtual humans,” in Intelligent Virtual Agents: 15th International Conference, IVA 2015, Delft, The Netherlands, August 26-28, 2015, Proceedings 15 . Springer, 2015, pp. 201–215

    2015

  9. [17]

    Negotiating globally: How to negotiate deals, resolve disputes, and make decisions across cultural boundaries,

    J. M. Brett, “Negotiating globally: How to negotiate deals, resolve disputes, and make decisions across cultural boundaries,” 2014

    2014

  10. [18]

    The interpersonal effects of anger and happiness in negotiations

    G. A. Van Kleef, C. K. De Dreu, and A. S. Manstead, “The interpersonal effects of anger and happiness in negotiations.” Journal of personality and social psychology , vol. 86, no. 1, p. 57, 2004

    2004

  11. [19]

    The effect of expression of anger and happiness in computer agents on negotiations with humans,

    C. M. de Melo, P. Carnevale, and J. Gratch, “The effect of expression of anger and happiness in computer agents on negotiations with humans,” in The 10th International Conference on Autonomous Agents and Multiagent Systems-Volume 3, 2011, pp. 937–944

    2011

  12. [20]

    Conflict escalation in organizations,

    D. G. Pruitt, “Conflict escalation in organizations,” in The psychology of conflict and conflict management in organizations . Psychology Press, 2007, pp. 261–282

    2007

  13. [21]

    Breaking the bonds of reciprocity in negotiations,

    J. M. Brett, D. L. Shapiro, and A. L. Lytle, “Breaking the bonds of reciprocity in negotiations,” Academy of Management Journal , vol. 41, no. 4, pp. 410–424, 1998

    1998

  14. [22]

    Group-based hatred in intractable conflict in israel,

    E. Halperin, “Group-based hatred in intractable conflict in israel,” Journal of Conflict resolution , vol. 52, no. 5, pp. 713–736, 2008

    2008

  15. [23]

    A survey on sentiment analysis methods, applications, and challenges,

    M. Wankhade, A. C. S. Rao, and C. Kulkarni, “A survey on sentiment analysis methods, applications, and challenges,” Artificial Intelligence Review, vol. 55, no. 7, pp. 5731–5780, 2022

    2022

  16. [24]

    Emotion recognition in conversation: Research challenges, datasets, and recent advances,

    S. Poria, N. Majumder, R. Mihalcea, and E. Hovy, “Emotion recognition in conversation: Research challenges, datasets, and recent advances,” IEEE access, vol. 7, pp. 100 943–100 953, 2019

    2019

  17. [25]

    To- wards emotion-aware agents for improved user satisfaction and partner perception in negotiation dialogues,

    K. Chawla, R. Clever, J. Ramirez, G. M. Lucas, and J. Gratch, “To- wards emotion-aware agents for improved user satisfaction and partner perception in negotiation dialogues,” IEEE Transactions on Affective Computing, 2023

    2023

  18. [26]

    Is gpt a computational model of emotion?

    A. N. Tak and J. Gratch, “Is gpt a computational model of emotion?” in 2023 11th International Conference on Affective Computing and Intelligent Interaction (ACII) , 2023, pp. 1–8

    2023

  19. [27]

    Investigating large language models’ perception of emotion using appraisal theory,

    N. Yongsatianchot, P. G. Torshizi, and S. Marsella, “Investigating large language models’ perception of emotion using appraisal theory,” in 2023 11th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) . IEEE, 2023, pp. 1–8

    2023

  20. [28]

    Evaluating subjective cognitive appraisals of emotions from large language models,

    H. Zhan, D. C. Ong, and J. J. Li, “Evaluating subjective cognitive appraisals of emotions from large language models,” arXiv preprint arXiv:2310.14389, 2023

    2023 arXiv

  21. [29]

    Fine-grained affective processing capabilities emerging from large language models,

    J. Broekens, B. Hilpert, S. Verberne, K. Baraka, P. Gebhard, and A. Plaat, “Fine-grained affective processing capabilities emerging from large language models,” in 2023 11th International Conference on Affective Computing and Intelligent Interaction (ACII) . IEEE, 2023, pp. 1–8

    2023

  22. [30]

    Emotional intelligence of large language models,

    X. Wang, X. Li, Z. Yin, Y . Wu, and J. Liu, “Emotional intelligence of large language models,” Journal of Pacific Rim Psychology , vol. 17, p. 18344909231213958, 2023

    2023

  23. [31]

    Casino: A corpus of campsite negotiation dialogues for automatic ne- gotiation systems,

    K. Chawla, J. Ramirez, R. Clever, G. Lucas, J. May, and J. Gratch, “Casino: A corpus of campsite negotiation dialogues for automatic ne- gotiation systems,” in Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Hu...

    2021

  24. [32]

    Lioness lab: a free web-based platform for conducting interactive experiments online,

    M. Giamattei, K. S. Yahosseini, S. G ¨achter, and L. Molleman, “Lioness lab: a free web-based platform for conducting interactive experiments online,” Journal of the Economic Science Association , vol. 6, no. 1, pp. 95–111, 2020

    2020

  25. [33]

    Information exchange in negotiation,

    L. L. Thompson, “Information exchange in negotiation,” Journal of experimental social psychology , vol. 27, no. 2, pp. 161–179, 1991

    1991

  26. [34]

    Learning about the opponent in automated bilateral negotiation: a comprehensive survey of opponent modeling techniques,

    T. Baarslag, M. J. Hendrikx, K. V . Hindriks, and C. M. Jonker, “Learning about the opponent in automated bilateral negotiation: a comprehensive survey of opponent modeling techniques,” Autonomous Agents and Multi-Agent Systems, vol. 30, pp. 849–898, 2016

    2016

  27. [35]

    An agenda-based framework for multi-issue negotiation,

    S. S. Fatima, M. Wooldridge, and N. R. Jennings, “An agenda-based framework for multi-issue negotiation,” Artificial Intelligence, vol. 152, no. 1, pp. 1–45, 2004

    2004

  28. [36]

    Iago: interactive arbitration guide online,

    J. Mell and J. Gratch, “Iago: interactive arbitration guide online,” in Proceedings of the 2016 International Conference on Autonomous Agents & Multiagent Systems , 2016, pp. 1510–1512

    2016

  29. [37]

    Deal or no deal? end-to-end learning of negotiation dialogues,

    M. Lewis, D. Yarats, Y . Dauphin, D. Parikh, and D. Batra, “Deal or no deal? end-to-end learning of negotiation dialogues,” in Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, 2017, pp. 2443–2453

    2017

  30. [38]

    Decoupling strategy and generation in negotiation dialogues,

    H. He, D. Chen, A. Balakrishnan, and P. Liang, “Decoupling strategy and generation in negotiation dialogues,” in Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing , E. Riloff, D. Chiang, J. Hockenmaier, and J. Tsujii, Eds. Brussels, Belgium...

    2018

  31. [39]

    Be selfish, but wisely: Investigating the impact of agent personality in mixed-motive human-agent interactions,

    K. Chawla, I. Wu, Y . Rong, G. Lucas, and J. Gratch, “Be selfish, but wisely: Investigating the impact of agent personality in mixed-motive human-agent interactions,” in Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing , H. Bouamor, J. Pin...

    2023

  32. [40]

    Agents that reason and negotiate by arguing,

    S. Parsons, C. Sierra, and N. Jennings, “Agents that reason and negotiate by arguing,” 1998

    1998

  33. [41]

    Spliddit: Unleashing fair division algorithms,

    J. Goldman and A. D. Procaccia, “Spliddit: Unleashing fair division algorithms,” ACM SIGecom Exchanges, vol. 13, no. 2, pp. 41–46, 2015

    2015

  34. [42]

    How emotions work: The social functions of emotional expression in negotiations,

    M. W. Morris and D. Keltner, “How emotions work: The social functions of emotional expression in negotiations,” Research in organizational behavior, vol. 22, pp. 1–50, 2000

    2000

  35. [43]

    Attitude and response to emotion in dispute resolution: The experience of mediators,

    K. Douglas and C. Coburn, “Attitude and response to emotion in dispute resolution: The experience of mediators,” Flinders LJ, vol. 16, p. 111, 2014

    2014

  36. [44]

    Are you angry (happy, sad) or aren’t you? emotion detection difficulty in email negotiation,

    C. Laubert and J. Parlamis, “Are you angry (happy, sad) or aren’t you? emotion detection difficulty in email negotiation,” Group Decision and Negotiation, vol. 28, pp. 377–413, 2019

    2019

  37. [45]

    An argument for basic emotions,

    P. Ekman, “An argument for basic emotions,” Cognition & emotion , vol. 6, no. 3-4, pp. 169–200, 1992

    1992

  38. [46]

    Emotional expressions reconsidered: Challenges to inferring emotion from human facial movements,

    L. F. Barrett, R. Adolphs, S. Marsella, A. M. Martinez, and S. D. Pollak, “Emotional expressions reconsidered: Challenges to inferring emotion from human facial movements,” Psychological science in the public interest, vol. 20, no. 1, pp. 1–68, 2019

    2019

  39. [47]

    Multilingual language models are not multicultural: A case study in emotion,

    S. Havaldar, S. Rai, B. Singhal, L. Liu, S. C. Guntuku, and L. Ungar, “Multilingual language models are not multicultural: A case study in emotion,” arXiv preprint arXiv:2307.01370 , 2023

    2023 arXiv

  40. [48]

    Can language model moderators improve the health of online discourse?

    H. Cho, S. Liu, T. Shi, D. Jain, B. Rizk, Y . Huang, Z. Lu, N. Wen, J. Gratch, E. Ferrara, and J. May, “Can language model moderators improve the health of online discourse?” in Proceedings of the 2024 Conference of the North American Chapter of the Association for Computation...

    2024

  41. [49]

    Negotiage: De- velopment and pilot testing of an artificial intelligence-based family caregiver negotiation program,

    A. Murawski, V . Ramirez-Zohfeld, J. Mell, M. Tschoe, A. Schierer, C. Olvera, J. Brett, J. Gratch, and L. A. Lindquist, “Negotiage: De- velopment and pilot testing of an artificial intelligence-based family caregiver negotiation program,” Journal of the American Geriatrics Soc...

    2024

  42. [50]

    Re- hearsal: Simulating conflict to teach conflict resolution,

    O. Shaikh, V . E. Chai, M. Gelfand, D. Yang, and M. S. Bernstein, “Re- hearsal: Simulating conflict to teach conflict resolution,” in Proceedings of the CHI Conference on Human Factors in Computing Systems , 2024, pp. 1–20

    2024

  43. [51]

    Kodis: A multicultural dispute resolution dialogue corpus,

    J. A. Hale, S. Rakshit, K. Chawla, J. M. Brett, and J. Gratch, “Kodis: A multicultural dispute resolution dialogue corpus,” in Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Asso- ciation for Computational Linguistics: Human Language Technolog...

    2025

  44. [52]

    Dignity, face, and honor cultures: A study of negotiation strategy and outcomes in three cultures,

    S. Aslani, J. Ramirez-Marin, J. Brett, J. Yao, Z. Semnani-Azad, Z.-X. Zhang, C. Tinsley, L. Weingart, and W. Adair, “Dignity, face, and honor cultures: A study of negotiation strategy and outcomes in three cultures,” Journal of Organizational Behavior, vol. 37, no. 8, pp. 1178...

    2016

  45. [53]

    The utility of relationships in ne- gotiation,

    A. D. Brown and J. R. Curhan, “The utility of relationships in ne- gotiation,” The Oxford handbook of economic conflict resolution , pp. 137–154, 2012

    2012

  46. [54]

    What do people value when they negotiate? mapping the domain of subjective value in negotiation

    J. R. Curhan, H. A. Elfenbein, and H. Xu, “What do people value when they negotiate? mapping the domain of subjective value in negotiation.” Journal of personality and social psychology , vol. 91, no. 3, p. 493, 2006

    2006

  47. [55]

    Are llms effective negotiators? systematic evaluation of the multifaceted capabilities of llms in negotiation dialogues,

    D. Kwon, E. Weiss, T. Kulshrestha, K. Chawla, G. M. Lucas, and J. Gratch, “Are llms effective negotiators? systematic evaluation of the multifaceted capabilities of llms in negotiation dialogues,” in Proceed- ings of the 2023 Conference on Empirical Methods in Natural Language...

    2023

  48. [56]

    The influence of anger and compassion on negotiation performance,

    K. G. Allred, J. S. Mallozzi, F. Matsui, and C. P. Raia, “The influence of anger and compassion on negotiation performance,” Organizational Behavior and Human Decision Processes, vol. 70, no. 3, pp. 175–187, 1997. [Online]. Available: https://www.sciencedirect.com/science/arti...

    1997

  49. [57]

    Linking Emotion to the Conflict Face-Negotiation Theory: A U.S.-China Investigation of the Mediating Effects of Anger, Compassion, and Guilt in Interpersonal Conflict,

    Q. Zhang, S. Ting-Toomey, and J. G. Oetzel, “Linking Emotion to the Conflict Face-Negotiation Theory: A U.S.-China Investigation of the Mediating Effects of Anger, Compassion, and Guilt in Interpersonal Conflict,” Human Communication Research, vol. 40, no. 3, pp. 373–395, 07 2...

    2014 doi

  50. [58]

    Gpt-4 technical report,

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat et al. , “Gpt-4 technical report,” arXiv preprint arXiv:2303.08774 , 2023

    2023 arXiv

  51. [59]

    Is” a helpful assistant

    M. Zheng, J. Pei, and D. Jurgens, “Is” a helpful assistant” the best role for large language models? a systematic evaluation of social roles in system prompts,” arXiv preprint arXiv:2311.10054 , 2023

    2023 arXiv

  52. [60]

    A survey on in-context learning,

    Q. Dong, L. Li, D. Dai, C. Zheng, Z. Wu, B. Chang, X. Sun, J. Xu, and Z. Sui, “A survey on in-context learning,” arXiv preprint arXiv:2301.00234, 2022

    2022 arXiv

  53. [61]

    A. R. Hochschild, The managed heart: Commercialization of human feeling. Univ of California Press, 2003

    2003

  54. [62]

    Emotional labor: The role of employee acting strategies on customer emotional experience and subsequent buying decisions,

    C. Tang, C. R. Seal, S. E. Naumann, and K. Miguel, “Emotional labor: The role of employee acting strategies on customer emotional experience and subsequent buying decisions,” International Review of Management and Marketing, vol. 3, no. 2, pp. 50–57, 2013

    2013

  55. [63]

    Deepseek-v3 technical report,

    A. Liu, B. Feng, B. Xue, B. Wang, B. Wu, C. Lu, C. Zhao, C. Deng, C. Zhang, C. Ruan et al., “Deepseek-v3 technical report,” arXiv preprint arXiv:2412.19437, 2024

    2024 arXiv

  56. [64]

    The llama 3 herd of models,

    A. Grattafiori, A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Vaughan et al. , “The llama 3 herd of models,” arXiv preprint arXiv:2407.21783 , 2024

    2024 arXiv

  57. [65]

    Gpt-4o mini: Advancing cost-efficient intelligence,

    OpenAI, “Gpt-4o mini: Advancing cost-efficient intelligence,” 2024. [Online]. Available: https://openai.com/index/gpt-4o-mini-advancing- cost-efficient-intelligence/

    2024

  58. [66]

    Rehearsal: Simulating conflict to teach conflict resolution,

    O. Shaikh, V . E. Chai, M. Gelfand, D. Yang, and M. S. Bernstein, “Rehearsal: Simulating conflict to teach conflict resolution,” in Proceedings of the CHI Conference on Human Factors in Computing Systems , ser. CHI ’24. New York, NY , USA: Association for Computing Machinery, ...

    2024

  59. [67]

    Ai-mediated dispute resolu- tion,

    J. Hale, H. Kim, A. Choi, and J. Gratch, “Ai-mediated dispute resolu- tion,” in Proceedings of the AAAI Symposium Series, vol. 5, no. 1, 2025, pp. 67–70

    2025

  60. [68]

    An application of large language models to coding negotiation transcripts,

    R. Friedman, J. Cho, J. Brett, X. Zhan, N. Han, S. Kannan, Y . Ma, J. Spencer-Smith, E. J ¨ackel, A. Zerres et al. , “An application of large language models to coding negotiation transcripts,” arXiv preprint arXiv:2407.21037, 2024

    2024 arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.