The DOI matches its Crossref record; 74 of the 74 references it cites were found in a registry or an index; 7 of the 19 sources its claims cite were checked for support.
"It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools
Open accessAbstract
While revolutionary AI-powered code generation tools have been rising rapidly, we know little about how and how to help software developers form appropriate trust in those AI tools. Through a two-phase formative study, we investigate how online communities shape developers’ trust in AI tools and how we can leverage community features to facilitate appropriate user trust. Through interviewing 17 developers, we find that developers collectively make sense of AI tools using the experiences shared by community members and leverage community signals to evaluate AI suggestions. We then surface design opportunities and conduct 11 design probe sessions to explore the design space of using community features to support user trust in AI code generation systems. We synthesize our findings and extend an existing model of user trust in AI technologies with sociotechnical factors. We map out the design considerations for integrating user community into the AI code generation experience.
Overview
CoChat Summary
## TL;DR
Online developer communities shape trust in AI code generation tools in two ways: by helping developers collectively make sense of the tools through shared experiences, and by giving them signals to judge the quality of AI suggestions. In two qualitative studies, the authors show that developers want community content that is vivid, realistic, and reproducible, but they also want it integrated into their workflow and protected for privacy.
## Key findings
- In Study 1, interviews with 17 developers showed two trust-building pathways: community-curated experiences and community-generated evaluation signals.
- Developers wanted experience posts with vivid descriptions/explanations, realistic programming tasks, diverse use cases, and project setup/dependency details; they also reported problems with missing details for reproducibility, lack of dedicated sharing channels, and the high cost of describing/sharing experiences.
- For evaluation, developers valued votes/ratings, discussion context, and identity signals such as contribution history, badges, and expertise; they also noted a trade-off between these signals and the productivity benefit of AI tools.
- In Study 2, 11 developers reacted to design probes for community evaluation signals and experience-sharing spaces; they liked the ideas but wanted the signals to be digestible at a glance, more contextual, and more clearly tied to privacy and consent.
## Methods
- Two-phase formative qualitative study centered on AI-powered code generation tools, especially GitHub Copilot.
- Study 1: semi-structured interviews with 17 developers, conducted online from July to August 2022, totaling 980 minutes of interview data; analyzed using inductive thematic analysis.
- Study 2: design probe sessions with 11 developers, conducted via recorded video calls in August 2022, totaling 646 minutes; analyzed using deductive thematic analysis over the probe features.
## Limitations
- The authors state the work is focused on AI-powered code generation tools such as Copilot, Tabnine, Kite, and Amazon CodeWhisperer, and cannot be claimed to apply to all AI applications or domains.
- Study 2 used Copilot-based stimuli and recruited participants familiar with Copilot, which may have excluded people unfamiliar with Copilot and may have biased thinking toward Copilot/VS Code design.
- The paper explicitly notes a gender imbalance in Study 1: 17 participants, with only two women and no gender-nonconforming participants; this may have biased the findings.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
1 Valid · 8 Not valid · 9 Not checked · 185 Own results and interpretations
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“a recent study has shown that trust is situated and shaped by users' information-seeking and assessment practices on online social platforms [ 72 ].”
§ "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools > 2.1 User Trust in AI and the MATCH Model
What the sources say
- Could not verifycites [72] 72. Albert Ziegler, Eirini Kalliamvakou, X. Alice Li, Andrew Rice, Devon Rifkin, Shawn Simister, Ganesh Sittampalam, and Edward Aftandilian. 2022. Productivity assessment of neural code completion. In Proceedings of the 6th ACM SIGPLAN International Symposium on Machine Programming (MAPS 2022) . Association for Computing Machinery, New York, NY, USA, 21-29. DOI: https://doi.org/10.1145/3520312.3534864 sourceExperimental — accuracy not yet measured
Section 4, “What Drives Perceived Productivity?” (page not available), examines developers’ GitHub Copilot usage and perceived productivity, not trust on online social platforms. Searches for situated trust, information-seeking, and assessment practices found no passage in our copy addressing the attributed claim. (LLM, conf=0.76)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
The cited sources that could be checked couldn’t confirm or dispute this claim.
In the paper
“Ehsan et al. [ 22 ] discovered that users' perceptions and needs in AI are situated in a social, collaborative setting.”
§ "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools > 2.1 User Trust in AI and the MATCH Model
What the sources say
- Could not verifycites [22] 22. Neil A. Ernst and Gabriele Bavota. 2022. AI-driven development is here: Should you worry? IEEE Software 39, 2 (2022), 106-110. DOI: https://doi.org/10.1109/MS.2021.3133805 sourceExperimental — accuracy not yet measured
The ‘Sociotechnical Questions’ passage (p. 4) discusses integrating an automated repair bot into a human team, but does not report a finding about users’ perceptions or needs in social, collaborative settings. Searches for the quoted claim and for Ehsan with its key terms found no such finding in our copy. (LLM, conf=0.73)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Specifically, developers can build a mental model of the capacity of the AI from others' experience, which includes setting reasonable expectations, learning strategies on when to trust the AI, forming empirical understanding of how AI generates suggestions, and developing awareness of broader implications of AI-generated code.”
§ "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools > 3.2.2 How Community-curated Experiences Help Developers Build Trust with AI Tools.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Our finding suggests that developers can build appropriate trust with AI code generation tools as a result of engaging with community-curated experience.”
§ "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools > 3.2.2 How Community-curated Experiences Help Developers Build Trust with AI Tools.
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We refer to such collections of sharing about community members' interaction with AI as community-curated experiences , which allow developers to form a mental model about the capacity of AI code generation tools and how they can use those tools, building trust in the process.”
§ "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools > 3.2 Finding 1: Building Trust with Community-curated Experiences
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
103 more claims
Create a free account to see all 203 claims and every source verdict.
- Probably not a substantial claim. 67
- It may state two separate claims as one. 49
- The quote only partly states this claim. 3
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“However, most of these studies focus on AI-assisted tool use in the context of understanding, generating, or modifying source code, providing little insight into the role AI plays in developer information seeking and its impact.”
From The Evolution of Information Seeking in Software Development: Understanding the Role and Impact of AI Assistants (2025) · cites this paper as [12]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Given the changing landscape in developer tooling, many studies have investigated the use and impact of AI-assisted tools on developer engagement and productivity [12,13,37,40].”
From The Evolution of Information Seeking in Software Development: Understanding the Role and Impact of AI Assistants (2025) · cites this paper as [12]
In the citing paper
“Prior work identifies multiple influences on developers' trust, including expectation-setting, workflow compatibility, and community-based cues such as peer experience and support [26, 116, 143].”
From What Needs Attention? Prioritizing Drivers of Developers' Trust and Adoption of Generative AI (2025) · cites this paper as [26]
In the citing paper
“Prior work shows that developers often place trust in AI based on its performance, presentation, and output quality (accuracy and consistency), which serve as proxies for a system's perceived credibility [26, 44, 143, 152].”
From What Needs Attention? Prioritizing Drivers of Developers' Trust and Adoption of Generative AI (2025) · cites this paper as [26]
Citations
Cites
Using thematic analysis in psychology
Braun, V., & Clarke, V. (2006) · Qualitative Research in Psychology · doi:10.1191/1478088706qp063oa · Cited by 194,414
Situated Learning: Legitimate Peripheral Participation
Lave, J., & Wenger, E. (1991) · Cambridge University Press eBooks · doi:10.1017/cbo9780511815355 · Cited by 35,324
Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective
Haraway, D. (1988) · Feminist Studies · doi:10.2307/3178066 · Cited by 18,364
Social network sites: definition, history, and scholarship
boyd, D. M., & Ellison, N. B. (2007) · Journal of Computer-Mediated Communication · doi:10.1111/j.1083-6101.2007.00393.x · Cited by 16,251
An integrative model of organizational trust
Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995) · The Academy of Management Review · doi:10.2307/258792 · Cited by 10,736 · cited 1×
“An AI model is typically built with certain trustworthiness attributes (leftmost box in the figure), such as its ability, benevolence, and integrity, that can directly affect users’ trust [ 19 , 48 , 68 , 69 ].”
BackgroundTrust in Automation: Designing for Appropriate Reliance
Lee, J. D., & See, K. A. (2004) · Human Factors: The Journal of the Human Factors and Ergonomics Society · doi:10.1518/hfes.46.1.50_30392 · Cited by 6,605 · cited 3×
“…may not be aware of cases where AI can fail, granting too much trust to AI and resulting in mistakes, similar to what we see in the literature on overreliance on and misuse of AI systems [ 21 , 43 , 51 ]; or developers may deem AI completely useless only based on their own few negative experiences.”
BackgroundResultGuidelines for Human-AI Interaction
Amershi, S., Weld, D., Vorvoreanu, M., Fourney, A., Nushi, B., Collisson, P., Suh, J., Iqbal, S., Bennett, P. N., Inkpen, K., et al. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300233 · Cited by 2,439 · cited 1×
“Trust has been highlighted as one of the most important factors in multiple design guidelines for human-AI collaboration [2, 15].”
BackgroundSchön, D. A. (2017) · doi:10.4324/9781315237473 · Cited by 1,740
The role of trust in automation reliance
Dzindolet, M. T., Peterson, S. A., Pomranky, R. A., Pierce, L. G., & Beck, H. P. (2003) · International Journal of Human-Computer Studies · doi:10.1016/s1071-5819(03)00038-7 · Cited by 1,276 · cited 2×
“A misalignment between the trust of users and AI’s ability could lead to user overreliance and misuse of AI systems, resulting in unwanted consequences [19, 36, 43].”
BackgroundResultProceedings of the 2021 CHI Conference on Human Factors in Computing Systems
(2021) · doi:10.1145/3411764 · Cited by 1,240
Research works printed but not linked: 3
- 10.1145/1145/3586030 · doi:10.1145/1145/3586030
- 10.5555/2207798 · doi:10.5555/2207798
- Samir Passi and Mihaela Vorvoreanu. 2022. Overreliance on AI literature review. (2022) 24.
Cited by
Navigating the Complexity of Generative AI Adoption in Software Engineering
Russo, D. (2024) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3652154 · Cited by 217
“’s [14] exploration of community influence on developers’ trust.”
BackgroundInvestigating and Designing for Trust in AI-powered Code Generation Tools
Wang, R., Cheng, R., Ford, D., & Zimmermann, T. (2024) · The 2024 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3630106.3658984 · Cited by 87
“[15], so that users could further extend their understanding of the capacities and limitations of AI and update their trust.”
BackgroundSelf-citation"I look at it as the king of knowledge": How Blind People Use and Understand Generative AI Tools
Adnin, R., & Das, M. (2024) · The 26th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663548.3675631 · Cited by 73
“Nowadays, people are incorporating GenAI tools (e.g., ChatGPT [73], Google Gemini [33], Microsoft Copilot [65], and Claude [5]) into a wide variety of domains, including education [36, 62, 83], programming [16, 54, 92], communication [14], and content creation [45, 56].”
BackgroundMethodGenerative AI Meets Knowledge Management: Insights From Software Development Practices
Kirchner, K., Bolisani, E., Kassaneh, T. C., Scarso, E., & Taraghi, N. (2025) · Knowledge and Process Management · doi:10.1002/kpm.70004 · Cited by 52
AI-Augmented Predictions: LLM Assistants Improve Human Forecasting Accuracy
Schoenegger, P., Park, P. S., Karger, E., Trott, S., & Tetlock, P. E. (2024) · ACM Transactions on Interactive Intelligent Systems · doi:10.1145/3707649 · Cited by 69
“…et al. 2023), summarization (Goyal, Li, and Durrett 2023), abstract categorization of objects (Atari et al. 2023), programming (Bubeck et al. 2023; Cheng et al. 2024), spear phishing cyber attacks (Hazell 2023; Heiding et al. 2023), human personality (Schoenegger et al. 2024a), robotics (Vemprala…”
Background
Metadata
- Authors
- Cheng, R., Wang, R., Zimmermann, T., & Ford, D.
- Affiliations
- University of Washington (US)
- Microsoft (United States) (US)
- Venue
- ACM Transactions on Interactive Intelligent Systems
- Published
- 2024-05-15
- Volume
- 14
- Issue
- 2
- Pages
- 1-39
- ISSN
- 2160-6455 (ISSN-L)
- 2160-6463
- OpenAlex
- W4392623181
- Semantic Scholar
- 3078a916b57e465614efea50628c889799e32289
- Cited by
- 47
- References
- 68
- Citations by year
- 2026: 21 · 2025: 20 · 2024: 6
- FWCI
- 17.73
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- ACM Transactions on Interactive Intelligent Systems (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Computer Science › Information Systems › Software Engineering Research
- Physical Sciences › Computer Science › Information Systems › Software Engineering Techniques and Practices
- Physical Sciences › Computer Science › Human-Computer Interaction › Usability and User Interface Design
- Keywords
- online developer communities, software developers
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-12, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-12 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools. https://app.cochat.ai/papers/10.1145/3651990
- The paper
- Cheng, R., Wang, R., Zimmermann, T., & Ford, D. (2024). "It would work for me too": How Online Communities Shape Software Developers' Trust in AI-Powered Code Generation Tools. ACM Transactions on Interactive Intelligent Systems. https://doi.org/10.1145/3651990
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.