Pith. sign in

REVIEW 4 major objections 5 minor 66 references

Enacting Constructive Conflicts with AI Agents to Enhance Reconsideration among Novice Interaction Designers

T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read An AI agent that enacts constructive conflict, by voicing stakeholder pushback and letting designers steer it, makes novice designers reconsider and revise their design proposals more than written prompts alone.

desk verdict A real framework-level result with an abstract that overclaims the agent's specific effects; the direct IE-vs-SG contrasts are missing where they matter most. read the letter →

arxiv 2608.04166 v1 pith:XZ5K5QE7 submitted 2026-08-04 cs.HC

Howard Ziyu Han , Nikolas Martelaro This is my paper · ORCID
classification cs.HC
keywords adversarialdesignconstructiveconflictAIagentnoviceinteractiondesignersstakeholdertensionrevisiondevil'sadvocatehuman-agent
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 reports a controlled experiment with 45 interaction design students who redesigned a civic 311 reporting website. Students who worked through a written constructive-conflict framework by themselves reported more self-reconsideration and made more edits than students who simply reviewed their own proposal. Students who instead interacted with an AI agent that enacted the same framework, generating stakeholder pushback and letting the designer steer the conflict, showed the same gains plus more revisions and deletions of their own ideas. The paper's central claim is that agent-enacted conflict converts reconsideration into concrete design actions and deepens engagement with divergent stakeholder perspectives, beyond what the framework alone achieves.

What carries the argument

The central object is an AI agent embedded in a Miro whiteboard that enacts constructive conflict. It collects the designer's context, then produces four pushback points that pair a specific stakeholder perspective with a challenge to the proposal. The designer steers the next round through point-level tagging (useful, not useful, custom) and stance-level framing using two 'Consensus' and 'Command' frames, and a second round narrows to directions the designer chose to probe. A sub-agent retrieves prior conflict examples and documents on agonistic pluralism to ground the pushback. The paper treats this interactive steering loop, not the conflict framework alone, as what turns awareness into action.

What would settle it

A direct test would compare Stepwise Guidance prompts that are matched to the agent's output on concreteness, specificity, and the number of stakeholder perspectives, and ask participants to rate how applicable each prompt was to their own proposal; if the gap between the two conditions disappears or reverses, the paper's attribution of the effect to interactive enactment would fail.

Watch

Extended reading notes

Core claim

The report argues that an antagonistic agent, built on adversarial design theory, can push novice designers past reflection into revision. In a between-subjects experiment, both the Stepwise Guidance and Interactive Engagement conditions scored higher on self-reconsideration than Self Reflection, and both made more improvements to their design proposals. The distinctive result is in idea editing: Interactive Engagement participants revised roughly 3.6 times more ideas than Self Reflection participants and were the only group to delete ideas, while Self Reflection participants never revised or deleted anything. The agent thus made the stakeholder costs of keeping an idea visible enough that novices cut, replaced, or rewrote parts of their proposals.

Load-bearing premise

The comparison that isolates the agent's effect assumes the Stepwise Guidance prompts presented the same conceptual content and imposed similar cognitive load as the agent's system prompt, so that the residual difference comes from enacting conflict interactively.

Editorial extensions

If this is right

  • If correct, design educators can use agent-enacted conflict to counter design fixation, since novices who interacted with the agent revised and deleted ideas while self-reflection groups did not.
  • The Stepwise Guidance condition's lower idea generation suggests that merely asking novices to think about stakeholder tensions can suppress new ideas unless the tensions are made concrete and interactive.
  • The agent's pushback surfaced stakeholder topics that neither baseline group considered on its own, implying the system can act as an interactive checklist for common civic concerns.
  • The steering mechanisms (tagging and Consensus/Command frames) were used by 14 of 15 participants, suggesting that making conflict steerable may be what keeps designers' confidence and agency intact.

Reading between the lines

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

  • A natural extension would test whether the effect transfers to professional designers or to later design stages, since the study only covers early ideation with novices in a single 15-minute intervention.
  • Because the paper notes synthetic pushback is not a substitute for real stakeholder input, grounding agent points in actual community input and measuring whether that changes how designers weigh the conflict would be a direct next step.
  • The descriptive pattern on self-efficacy, with Engagement trending higher than Guidance, hints that interactive steering may buffer the confidence costs of adversarial feedback, but that claim needs replication with a larger sample before being acted on.
  • Whether the agent's effects come from its interactivity or from the concreteness of its output is not settled by this design; a condition with pre-written concrete pushback without interactive steering would separate these.
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 / 5 minor

Summary. This paper reports a between-subjects experiment (N = 45 interaction design students) on a 311-reporting-website redesign task, comparing three conditions: Self Reflection (unsupported review), Stepwise Guidance (written prompts derived from the agent's system prompt), and Interactive Engagement (an LLM-based antagonistic agent in Miro that generates stakeholder pushback and lets participants steer a second round). The authors find that both framework conditions produce significantly higher self-reported reconsideration and more idea edits than Self Reflection, and that Interactive Engagement shows more revisions and larger self-reported shifts in design thinking. The paper argues that an antagonistic agent can turn reconsideration into concrete design actions and deepen engagement with divergent stakeholder perspectives, while also cautioning that synthetic pushback is not a substitute for real stakeholder input.

Significance. The contribution is potentially valuable. It extends devil's-advocate HAI from convergent decision tasks to open-ended interaction design, grounds the agent design in expert interviews and agonistic pluralism, and uses a three-arm design that attempts to separate a written conflict framework from an interactive agent. The paper is also transparent about limitations, reports effect sizes alongside Bonferroni-corrected pairwise tests, and states that the codebase will be released upon acceptance. The core contrast between structured conflict scaffolding and unsupported reflection is credible and well supported by the reported data. The distinctive claim, however, that agent enactment rather than the framework itself drives concrete design actions, is not yet established by the reported analyses; direct Interactive-Engagement-versus-Stepwise-Guidance tests are missing for key outcomes and one headline claim is contradicted by a non-significant contrast.

major comments (4)
  1. [Abstract; §4.1.1; §4.1.3] The abstract and conclusion overstate the agent-specific behavioral effects. For idea addition, the Interactive Engagement versus Stepwise Guidance contrast is explicitly non-significant (p_adj = .346, §4.1.1), so the claim that Interactive Engagement participants 'generated ... more ideas' is not supported. For deletions, §4.1.3 reports only descriptive counts (3 participants deleting 4 ideas in Interactive Engagement, 2 participants deleting 3 ideas in Stepwise Guidance, none in Self Reflection) with no inferential test, so the claim that they 'discarded more ideas' is likewise unsupported. The conclusion's statement that 'the agent produced effects neither baseline nor conceptual content achieved alone' (§7) is therefore stronger than the evidence. This is load-bearing because the paper's unique contribution is the enactment contrast, not the framework-versus-control comparison.
  2. [Introduction; §3.3.2; Appendix E] The enabling assumption that the Stepwise Guidance prompts 'mirror' the agent's system prompt is asserted but not validated. The Introduction and §3.3.2 state that the prompts mirror the system prompt and point to Appendix E for the correspondence, but no evidence is given that the two conditions present the same conceptual content, specificity, or cognitive load, or that Stepwise participants could apply the written prompts to their own proposals as readily as Interactive Engagement participants could react to situated pushback. If the written guidance was vaguer, more abstract, or harder to apply, then the observed differences between Interactive Engagement and Stepwise Guidance on revisions and self-reported design thinking could be due to content quality rather than to agent enactment. The authors should provide a content-overlap analysis or a manipulation check, and at minimum report the direct Interactive-Engagement-versus-Stepwise-Guidance comparison for every outcome that feeds the agent-enactment claim.
  3. [§3.3.3; §3.4.1] The behavioral counts underlying 'concrete design actions' rest on single-coder coding. §3.4.1 states 'The first author coded all design proposals,' and no inter-rater reliability statistic is reported anywhere in the manuscript. §3.3.3 says ambiguous cases were 'resolved by team discussion,' but this does not quantify agreement. Because additions, edits, revisions, and deletions are central to the headline behavioral claims, the paper should include a second coder and report agreement (for example, Cohen's kappa or proportional agreement) on the full set or a randomly selected subset of proposals.
  4. [§4.1.2; §4.2.4] Direct Interactive-Engagement-versus-Stepwise-Guidance pairwise tests are missing for the outcomes where the agent appears most differentiated. For 'the activity changed your design thinking,' the paper reports Interactive Engagement versus Self Reflection (p_adj = .003) and Stepwise Guidance versus Self Reflection (p_adj = .085) but never the direct Interactive Engagement versus Stepwise Guidance contrast. For total idea edits, §4.1.2 states only that the comparison 'was not significant' without reporting the p value or effect size; for revisions and enhancements, the Interactive-Engagement-versus-Stepwise-Guidance contrasts are likewise not reported. Without the full 3x3 pairwise matrix, the reader cannot determine whether any Interactive-Engagement-versus-Self-Reflection effect exceeds the framework-only condition. Report all pairwise comparisons with effect sizes and confidence intervals.
minor comments (5)
  1. [§5.1] The phrase 'dedevil's advocate gents' appears to be a typo for 'devil's advocate agents' and should be corrected.
  2. [References] Reference [4] contains malformed BibTeX-like fields ('prefix=van der useprefix=true family=Velde'), and reference [18] has an incomplete title beginning 'WHAT CAN WE LEARN FROM DESIGN.' These entries should be cleaned up.
  3. [Figure 3] The significance stars in Figure 3 should identify which pairwise contrasts are being tested; with three conditions and several outcome categories, a single star placement is ambiguous.
  4. [§4.3.1] The sentence beginning 'This implies a predominantly positive but discerning reception' generalizes from descriptive tag percentages without an inferential test; it should be softened or explicitly labeled as descriptive.
  5. [Appendix F; §4.2] The survey items are grouped into factors, but the manuscript does not state whether each reported analysis used a single item or a composite scale, nor does it report scale reliability. Clarify this for each dependent measure.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: empirical between-subjects comparison with independent behavioral and self-report measures; no fitted parameter renamed as prediction.

full rationale

This paper makes no mathematical or definitional derivation that reduces to its own inputs. The central claim—that agent-enacted constructive conflict increases reconsideration and concrete design actions—is supported by a three-condition between-subjects experiment with independently measured behavioral counts (idea additions, edits, deletions) and self-report scales. Neither the intervention conditions nor the outcome measures are defined in terms of one another. The Stepwise Guidance condition is described as using prompts that 'mirror' the agent's system prompt, but this is an asserted design property, not a fitted parameter; even if the mirroring were imperfect, that would be an internal-validity concern, not circularity. There are no fitted parameters renamed as predictions, no uniqueness theorem imported from the authors' prior work, and no ansatz smuggled in via self-citation. Self-citations in related work (e.g., reference [27] including co-author Martelaro) are contextual and not load-bearing for the reported results. The paper also explicitly acknowledges exploratory limitations and inflated effect-size estimates, which further indicates the claims are empirical rather than definitional. No circular step can be quoted or exhibited, so the circularity score is 0.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

No mathematical free parameters appear because the paper is an empirical study. The entries below are hand-chosen implementation parameters and background assumptions that the measured effects may depend on. Listing them makes explicit that the results are tied to one specific agent configuration.

free parameters (3)
  • Number of pushback points per round = 4 (fixed per round)
    Set by hand after pilot testing to balance breadth against cognitive load; the amount of conflict participants saw, and thus the observed effects, could depend on this count.
  • Generative model for pushback = GPT-4.1
    Chosen after pilot ratings because of faster generation; different models or prompt styles may produce different pushback quality and different user responses.
  • Adversarial tone in system prompt = aggressive and confrontational, no suggestions
    Hand-authored to give pushback 'bite'; results may not generalize to other tones or to agents that mix critique with suggestions.
assumptions (3)
  • domain assumption LLM-generated stakeholder pushback is a reasonable proxy for real stakeholder concerns.
    The agent's influence is interpreted as 'engagement with divergent stakeholder perspectives.' The authors themselves caution that synthetic points are not a substitute for real stakeholder input (Discussion, Section 5.2), so the central interpretation rests on this proxy assumption.
  • domain assumption Stepwise Guidance prompts and the agent's system prompt deliver the same conceptual framework.
    The comparison between Guidance and Engagement is used to isolate interactivity. The paper says the prompts 'mirror' the system prompt (Introduction; Section 3.3.2), but mirroring is not equivalence, and any content mismatch would confound the contrast.
  • domain assumption Self-report Likert items and researcher coding of sticky notes capture reconsideration and design improvement.
    The quantitative outcome variables are self-reported design thinking and single-coder idea classifications; the validity of the conclusions depends on these operationalizations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enacting Constructive Conflicts with AI Agents to Enhance Reconsideration among Novice Interaction Designers." pith.science (2026). https://pith.science/paper/XZ5K5QE7

@misc{pith2026260804166,
  author       = {Pith},
  title        = {Pith review of: Enacting Constructive Conflicts with AI Agents to Enhance Reconsideration among Novice Interaction Designers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XZ5K5QE7}},
  note         = {Machine review of arXiv:2608.04166}
}
read the original abstract

Generative AI agents are increasingly used in interaction design to facilitate ideation and offer critique, often following their own internal reasoning. These interactions tend to add design ideas and expand the design space. Our work explores an antagonistic role for design agents, prompting designers to engage with stakeholder tension. We built an AI agent inspired by adversarial design theory that enacts constructive conflict. We examine the agent's influence in a between-subjects experiment with 45 design students across three conditions: Self Reflection (unsupported review of the design proposal), Stepwise Guidance (written prompts that walk designers through a constructive-conflict framework), and Interactive Engagement (an AI agent that enacts the constructive-conflict framework interactively by synthesizing stakeholder pushback). The latter two conditions share the framework but differ in whether it is self-enacted or agent-enacted. Results show that, compared with Self Reflection, both the Stepwise Guidance and Interactive Engagement groups reported significantly higher self-reconsideration and made more improvements to their design proposals. Compared with Stepwise Guidance, the antagonistic agent introduced more conflictual perspectives, and participants in the Interactive Engagement condition generated and discarded more ideas. These findings suggest that agent-enacted constructive conflict can turn reconsideration into concrete design actions and deepen engagement with divergent stakeholder perspectives.

Figures

Figures reproduced from arXiv: 2608.04166 by the authors.

Figure 1
Figure 1. The Interactive Engagement condition’s Miro board, showing (left to right) the design challenge brief, the participant’s [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. The figure shows how ideas changed between itera [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Participants’ ratings for the Agent-specific post [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: The public interaction design challenge we used for the controlled study and the experts’ interview (mask city and [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: The instructions for how to interact with the agent [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: The stepwise instruction on how to think about constructive conflicts for the Stepwise Guidance group [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 51 canonical work pages

  1. [1]

    Dan Ariely and Michael I Norton. 2011. From thinking too little to thinking too much: a continuum of decision making.Wiley Interdisciplinary Reviews: Cognitive Science2, 1 (2011), 39–46

    work page 2011

  2. [2]

    Mohammadmehdi Ataei, Hyunmin Cheong, Daniele Grandi, Ye Wang, Nigel Morris, and Alexander Tessier. 2024. Elicitron: An LLM Agent-Based Simulation Framework for Design Requirements Elicitation. doi:10.48550/arXiv.2404.16045 arXiv:2404.16045 [cs]

    arXiv doi:10.48550/arxiv.2404.16045 2024

  3. [3]

    Erling Björgvinsson, Pelle Ehn, and Per-Anders Hillgren. 2012. Agonistic partici- patory design: working with marginalised social movements.CoDesign8, 2-3 (2012), 127–144

    work page 2012

  4. [4]

    Simone Borsci, Alessio Malizia, Martin Schmettow, prefix=van der useprefix=true family=Velde, given=Frank, Gunay Tariverdiyeva, Divyaa Balaji, and Alan Cham- berlain. 2022. The Chatbot Usability Scale: The Design and Pilot of a Usability Scale for Interaction with AI-Based Conversational Agents. 26, 1 (2022), 95–119. doi:10.1007/s00779-021-01582-9

    work page doi:10.1007/s00779-021-01582-9 2022

  5. [5]

    Richard Buchanan. 1992. Wicked problems in design thinking.Design issues8, 2 (1992), 5–21

    work page 1992

  6. [6]

    Zana Buçinca, Maja Barbara Malaya, and Krzysztof Z Gajos. 2021. To trust or to think: cognitive forcing functions can reduce overreliance on AI in AI-assisted decision-making.Proceedings of the ACM on Human-computer Interaction5, CSCW1 (2021), 1–21

    2021

  7. [7]

    Alice Cai, Ian Arawjo, and Elena L Glassman. 2024. Antagonistic AI.arXiv preprint arXiv:2402.07350(2024)

    arXiv 2024

  8. [8]

    Gilad Chen, Stanley M Gully, and Dov Eden. 2001. Validation of a new general self-efficacy scale.Organizational research methods4, 1 (2001), 62–83

    work page 2001

Show all 66 references
  1. [9]

    Ming-Huei Chen. 2006. Understanding the benefits and detriments of conflict on team creativity process.Creativity and innovation management15, 1 (2006), 105–116

    2006

  2. [10]

    Balanced

    XinHui Chen, Xiang Yuan, Hui Zhang, Ruixiao Zheng, and Wanyi Wei. 2025. Maintaining" Balanced" Conflict: Proactive Intervention Strategies of AI Voice Agents in Online Collaboration of Temporary Design Teams. InProceedings of the 2025 CHI Conference on Human Factors in Computi...

    2025

  3. [11]

    Myra Cheng, Cinoo Lee, Pranav Khadpe, Sunny Yu, Dyllan Han, and Dan Jurafsky

  4. [12]

    Erin Cherry and Celine Latulipe. 2014. Quantifying the creativity support of digital tools through the creativity support index.ACM Transactions on Computer- Human Interaction (TOCHI)21, 4 (2014), 1–25

    2014

  5. [13]

    Chun-Wei Chiang, Zhuoran Lu, Zhuoyan Li, and Ming Yin. 2024. Enhancing AI-Assisted Group Decision Making through LLM-Powered Devil’s Advocate. InProceedings of the 29th International Conference on Intelligent User Interfaces (IUI ’24). Association for Computing Machinery, New ...

    2024

  6. [14]

    Li-Yuan Chiou, Peng-Kai Hung, Rung-Huei Liang, and Chun-Teng Wang. 2023. Designing with AI: An Exploration of Co-Ideation with Image Generators. In Proceedings of the 2023 ACM Designing Interactive Systems Conference (DIS ’23). Association for Computing Machinery, New York, NY...

    2023

  7. [15]

    Leah Chong, Ayush Raina, Kosa Goucher-Lambert, Kenneth Kotovsky, and Jonathan Cagan. 2023. The Evolution and Impact of Human Confidence in Artificial Intelligence and in Themselves on AI-Assisted Decision-Making in Design.Journal of Mechanical Design145, 3 (2023), 031401. doi:...

    2023 doi

  8. [16]

    Nazli Cila. 2022. Designing Human-Agent Collaborations: Commitment, respon- siveness, and support. InProceedings of the 2022 CHI Conference on Human Factors in Computing Systems (CHI ’22). Association for Computing Machinery, New York, NY, USA, 1–18. doi:10.1145/3491102.3517500

    2022

  9. [17]

    Peter Dalsgaard. 2012. Participatory Design in Large-Scale Public Projects: Challenges and Opportunities. 28, 3 (2012), 34–47. jstor:23273837 https: //www.jstor.org/stable/23273837

    2012

  10. [18]

    WHAT CAN WE LEARN FROM DESIGN. 2006. What we learn when we engage in design: Implications for assessing design research.Educational design research (2006), 156

    2006

  11. [19]

    Carl DiSalvo. 2010. Design, democracy and agonistic pluralism. (2010)

    2010

  12. [20]

    Lynn Dombrowski, Ellie Harmon, and Sarah Fox. 2016. Social Justice-Oriented Interaction Design: Outlining Key Design Strategies and Commitments. InPro- ceedings of the 2016 ACM Conference on Designing Interactive Systems (DIS ’16). Association for Computing Machinery, New York...

    2016

  13. [21]

    Steven P Dow, Alana Glassco, Jonathan Kass, Melissa Schwarz, Daniel L Schwartz, and Scott R Klemmer. 2010. Parallel prototyping leads to better design results, more divergence, and increased self-efficacy.ACM Transactions on Computer- Human Interaction (TOCHI)17, 4 (2010), 1–24

    2010

  14. [22]

    Peitong Duan, Chin-Yi Cheng, Gang Li, Bjoern Hartmann, and Yang Li. 2024. UICrit: Enhancing Automated Design Evaluation with a UI Critique Dataset. In Proceedings of the 37th Annual ACM Symposium on User Interface Software and Technology. 1–17

    2024

  15. [23]

    Jennifer Fereday and Eimear Muir-Cochrane. 2006. Demonstrating rigor using thematic analysis: A hybrid approach of inductive and deductive coding and theme development.International journal of qualitative methods5, 1 (2006), 80–92

    2006

  16. [24]

    Gerhard Fischer, Kumiyo Nakakoji, Jonathan Ostwald, Gerry Stahl, and Tamara Sumner. 1993. Embedding critics in design environments.The knowledge engi- neering review8, 4 (1993), 285–307

    1993

  17. [25]

    Aakash Gautam, Deborah Tatar, Christian Matheis, and Gopinaath Kannabiran

  18. [26]

    Amanda Anne Geppert and Laura Ellen Forlano. 2022. Design for equivalence: Agonism for collective emancipation in participatory design. InProceedings of the Participatory Design Conference 2022-Volume 1. 158–168

    2022

  19. [27]

    Frederic Gmeiner, Kaitao Luo, Ye Wang, Kenneth Holstein, and Nikolas Martelaro

  20. [28]

    Heng Gu, Senthil Chandrasegaran, and Peter Lloyd. 2025. Synthetic Users: Insights from Designers’ Interactions with Persona-Based Chatbots.AI EDAM 39 (2025). doi:10.1017/S0890060424000283

    2025 doi

  21. [29]

    John Hattie and Helen Timperley. 2007. The Power of Feedback.Review of Educational Research77, 1 (2007), 81–112. doi:10.3102/003465430298487

    2007 doi

  22. [30]

    Theodore T Herbert and Ralph W Estes. 1977. Improving executive decisions by formalizing dissent: The corporate devil’s advocate.Academy of Management Review2, 4 (1977), 662–667

    1977

  23. [31]

    Perttu Hämäläinen, Mikke Tavast, and Anton Kunnari. 2023. Evaluating Large Language Models in Generating Synthetic HCI Research Data: A Case Study. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (CHI ’23). Association for Computing Machinery, N...

    2023

  24. [32]

    Jamieson, Piercarlo Valdesolo, and Brett J

    Jeremy P. Jamieson, Piercarlo Valdesolo, and Brett J. Peters. 2014. Sympathy for the devil? The physiological and psychological effects of being an agent (and target) of dissent during intragroup conflict.Journal of Experimental Social Psychology55 (Nov. 2014), 221–227. doi:10...

    2014 doi

  25. [33]

    Jansson and Steven M

    David G. Jansson and Steven M. Smith. 1991. Design Fixation.Design Studies12, 1 (1991), 3–11. doi:10.1016/0142-694X(91)90003-F

    1991 doi

  26. [34]

    Kushal Karwa. 2023. AI-powered career coaching: Evaluating feedback tools for design students.Indian Journal of Economics & Business. https://www. ash- winanokha. com/ijeb-v22-4-2023. php(2023)

    2023

  27. [35]

    Travis Kriplean, Jonathan Morgan, Deen Freelon, Alan Borning, and Lance Ben- nett. 2012. Supporting Reflective Public Thought with Considerit. InProceedings of the ACM 2012 Conference on Computer Supported Cooperative Work(Seattle Washington USA, 2012-02-11). ACM, 265–274. doi...

    2012

  28. [36]

    Tian Liang, Zhiwei He, Wenxiang Jiao, Xing Wang, Yan Wang, Rui Wang, Yujiu Yang, Shuming Shi, and Zhaopeng Tu. 2024. Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate. doi:10.48550/arXiv.2305. 19118 arXiv:2305.19118 [cs]

    2024 doi arXiv

  29. [37]

    Yiren Liu, Pranav Sharma, Mehul Oswal, Haijun Xia, and Yun Huang. 2025. Per- sonaFlow: Designing LLM-Simulated Expert Perspectives for Enhanced Research Ideation. InProceedings of the 2025 ACM Designing Interactive Systems Conference (New York, NY, USA, 2025-07-04)(DIS ’25). A...

    2025

  30. [38]

    Tao Long, Kendra Wannamaker, Jo Vermeulen, George Fitzmaurice, and Justin Matejka. 2025. FeedQUAC: Quick Unobtrusive AI-Generated Commentary. In Proceedings of the 13th International Conference on Human-Agent Interaction (HAI ’25). Association for Computing Machinery, New York...

    2025

  31. [39]

    Markham, Stephen G

    Stephen K. Markham, Stephen G. Green, and Raja Basu. 1991. Champions and Antagonists: Relationships with R&d Project Characteristics and Management. 8, 3 (1991), 217–242. doi:10.1016/0923-4748(91)90012-G

    1991 doi

  32. [40]

    Chantal Mouffe. 1999. Deliberative democracy or agonistic pluralism?Social research(1999), 745–758

    1999

  33. [41]

    Anwesha Mukherjee, Vagner Figueredo De Santana, and Alexis Baria. 2023. ImpactBot: Chatbot Leveraging Language Models to Automate Feedback and Promote Critical Thinking Around Impact Statements. InExtended Abstracts of the 2023 CHI Conference on Human Factors in Computing Syst...

    2023

  34. [42]

    Muller and Sarah Kuhn

    Michael J. Muller and Sarah Kuhn. 1993. Participatory design.Commun. ACM 36, 6 (June 1993), 24–28. doi:10.1145/153571.255960

    1993

  35. [43]

    Arisa Naka, Mamoru Yoshizoe, and Hiromitsu Hattori. 2025. Exploring Creative Human-Agent Discussion via Assigned Perspectives. InProceedings of the 13th In- ternational Conference on Human-Agent Interaction (HAI ’25). Association for Com- puting Machinery, New York, NY, USA, 1...

    2025

  36. [44]

    W Lawrence Neeley Jr, Kirsten Lim, April Zhu, and Maria C Yang. 2013. Building fast to think faster: exploiting rapid prototyping to accelerate ideation during early stage design. Ininternational design engineering technical conferences and computers and information in enginee...

    2013

  37. [45]

    Charlan Nemeth, Keith Brown, and John Rogers. 2001. Devil’s advocate versus authentic dissent: Stimulating quantity and quality.European Journal of Social Psychology31, 6 (2001), 707–720

    2001

  38. [46]

    Maria Adriana Neroni and Nathan Crilly. 2019. Whose Ideas Are Most Fixating, Your Own or Other People’s? The Effect of Idea Agency on Subsequent Design Behaviour.Design Studies60 (2019), 180–212. doi:10.1016/j.destud.2018.05.004

    2019 doi

  39. [47]

    Nicol and Debra Macfarlane-Dick

    David J. Nicol and Debra Macfarlane-Dick. 2006. Formative Assessment and Self- Regulated Learning: A Model and Seven Principles of Good Feedback Practice. Studies in Higher Education31, 2 (2006), 199–218. doi:10.1080/03075070600572090

    2006 doi

  40. [48]

    Changhoon Oh, Jungwoo Song, Jinhan Choi, Seonghyeon Kim, Sungwoo Lee, and Bongwon Suh. 2018. I lead, you help but only with enough details: Understanding user experience of co-creation with artificial intelligence. InProceedings of the 2018 CHI conference on human factors in c...

    2018

  41. [49]

    Alexi Orchard and Marcel O’Gorman. 2024. Fostering responsible innovation with critical design methods.Journal of Responsible Innovation11, 1 (2024), 2318823

    2024

  42. [50]

    Simon Parker and Martin Parker. 2017. Antagonism, Accommodation and Ago- nism in Critical Management Studies: Alternative Organizations as Allies. 70, 11 (2017), 1366–1387. doi:10.1177/0018726717696135

    2017 doi

  43. [51]

    Mirjana Prpa, Giovanni Maria Troiano, Matthew Wood, and Yvonne Coady. 2024. Challenges and Opportunities of LLM-Based Synthetic Personae and Data in HCI. InExtended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA ’24). Association for Computing Ma...

    2024

  44. [52]

    I Wayan Redhana and Ida Bagus Nyoman Sudria. 2020. Validity and reliability of critical thinking disposition inventory. In3rd International Conference on Innovative Research Across Disciplines (ICIRAD 2019). Atlantis Press, 282–287

    2020

  45. [53]

    David M Schweiger, William R Sandberg, and James W Ragan. 1986. Group approaches for improving strategic decision making: A comparative analysis of dialectical inquiry, devil’s advocacy, and consensus.Academy of management Journal29, 1 (1986), 51–71

    1986

  46. [54]

    Bernstein

    Omar Shaikh, Valentino Emil Chai, Michele Gelfand, Diyi Yang, and Michael S. Bernstein. 2024. Rehearsal: Simulating Conflict to Teach Conflict Resolution. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (New York, NY, USA, 2024-05-11)(CHI ’24). ...

    2024

  47. [55]

    Donghoon Shin, Soomin Kim, Ruoxi Shang, Joonhwan Lee, and Gary Hsieh

  48. [56]

    Simon Thuillard, M Adams, G Jelmini, S Schmutz, Andreas Sonderegger, and J Sauer. 2022. When humans and computers induce social stress through negative feedback: Effects on performance and subjective state.Computers in Human Behavior133 (2022), 107270

    2022

  49. [57]

    Matteo Vignoli, Clio Dosi, and Bernardo Balboni. 2023. Design thinking mindset: scale development and validation.Studies in Higher Education48, 6 (2023), 926– 940

    2023

  50. [58]

    Wen-Fan Wang, Chien-Ting Lu, Nil Ponsa i Campanyà, Bing-Yu Chen, and Mike Y. Chen. 2025. AIdeation: Designing a Human-AI Collaborative Ideation System for Concept Designers. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI ’25). Association f...

    2025

  51. [59]

    You’ve Got a Friend in Me

    Joel Wester, Andreas Brännström, Juan Carlos Nieves, and Niels van Berkel. 2023. “You’ve Got a Friend in Me”: A Formal Understanding of the Critical Friend Agent. InProceedings of the 11th International Conference on Human-Agent Interaction (HAI ’23). Association for Computing...

    2023

  52. [60]

    David C Wynn and Claudia M Eckert. 2017. Perspectives on iteration in design and development.Research in Engineering Design28, 2 (2017), 153–184

    2017

  53. [61]

    Xiaotong Xu, Arina Konnova, Bianca Gao, Cindy Peng, Dave Vo, and Steven P Dow. 2025. Productive vs. Reflective: How Different Ways of Integrating AI into Design Workflows Affect Cognition and Motivation. InProceedings of the 2025 CHI Conference on Human Factors in Computing Sy...

    2025

  54. [62]

    [specific group of people/community members] might push back: [A genuine objection, concern, or pushback they would voice - can be phrased as a question or statement]

    Yu-Chun Grace Yen, Isabelle Yan Pan, Grace Lin, Mingyi Li, Hyoungwook Jin, Mengyi Chen, Haijun Xia, Steven P Dow, et al. 2024. When to Give Feedback: Exploring Tradeoffs in the Timing of Design Feedback. (2024). 10 Appendix A Design Challenge Example Figure 5: The public inter...

    2024

  55. [2023]

    InProceedings of the 2023 CHI Conference on Human Factors in Computing Systems

    IntroBot: Exploring the Use of Chatbot-assisted Familiarization in Online Collaborative Groups. InProceedings of the 2023 CHI Conference on Human Factors in Computing Systems. ACM, Hamburg Germany, 1–13. doi:10.1145/3544548. 3580930

    2023 doi

  56. [2024]

    InProceedings of the Participatory Design Conference 2024: Exploratory Papers and Workshops-Volume 2

    Surfacing Conflicts in Participatory Design: Methodological Considerations. InProceedings of the Participatory Design Conference 2024: Exploratory Papers and Workshops-Volume 2. 194–197

    2024

  57. [2025]

    InProceedings of the 2025 ACM Designing Interactive Systems Conference

    Exploring the Potential of Metacognitive Support Agents for Human-AI Co- Creation. InProceedings of the 2025 ACM Designing Interactive Systems Conference. 1244–1269

    2025

  58. [2026]

    Science391, 6792 (2026)

    Sycophantic AI decreases prosocial intentions and promotes dependence. Science391, 6792 (2026). doi:10.1126/science.aec8352

    2026 doi

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.