Does More Advice Help? The Effects of Second Opinions in AI-Assisted Decision Making
Open accessThe DOI matches its Crossref record; 48 of the 48 references it cites were found in a registry or an index.
Abstract
AI assistance in decision-making has become popular, yet people's inappropriate reliance on AI often leads to unsatisfactory human-AI collaboration performance. In this paper, through three pre-registered, randomized human subject experiments, we explore whether and how the provision of second opinions may affect decision-makers' behavior and performance in AI-assisted decision-making. We find that if both the AI model's decision recommendation and a second opinion are always presented together, decision-makers reduce their over-reliance on AI while increase their under-reliance on AI, regardless whether the second opinion is generated by a peer or another AI model. However, if decision-makers have the control to decide when to solicit a peer's second opinion, we find that their active solicitations of second opinions have the potential to mitigate over-reliance on AI without inducing increased under-reliance in some cases. We conclude by discussing the implications of our findings for promoting effective human-AI collaborations in decision-making.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Cognitive Reflection and Decision Making
Frederick, S. (2005) · Journal of Economic Perspectives · doi:10.1257/089533005775196732 · Cited by 5,256
Algorithm aversion: People erroneously avoid algorithms after seeing them err.
Dietvorst, B. J., Simmons, J. P., & Massey, C. (2015) · Journal of Experimental Psychology: General · doi:10.1037/xge0000033 · Cited by 3,232 · cited 3×
“Interestingly, both the phenomenon of “algorithm aversion” (i.e., recommendations from humans are used more than those from AI) [21, 24, 111] and “algorithm appreciation” (i.e., recommendations from AI models are used more than those from humans) [55] are observed in different contexts.”
BackgroundAlgorithm appreciation: People prefer algorithmic to human judgment
Logg, J. M., Minson, J. A., & Moore, D. A. (2019) · Organizational Behavior and Human Decision Processes · doi:10.1016/j.obhdp.2018.12.005 · Cited by 2,061 · cited 2×
“Interestingly, both the phenomenon of “algorithm aversion” (i.e., recommendations from humans are used more than those from AI) [21, 24, 111] and “algorithm appreciation” (i.e., recommendations from AI models are used more than those from humans) [55] are observed in different contexts.”
Buçinca, Z., Malaya, M. B., & Gajos, K. Z. (2021) · arXiv · doi:10.48550/arxiv.2102.09692 · Cited by 1,124
Zhang, Y., Liao, Q. V., & Bellamy, R. K. E. (2020) · Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency · doi:10.1145/3351095.3372852 · Cited by 895
A survey of human-in-the-loop for machine learning
Wu, X., Xiao, L., Sun, Y., Zhang, J., Ma, T., & He, L. (2022) · Future Generation Computer Systems · doi:10.1016/j.future.2022.05.014 · Cited by 800 · cited 1×
“These scenarios can often been found in AI-assisted human labeling [1, 19, 20] and human-in-the-loop machine learning pipelines [63, 64, 104].”
Does the Whole Exceed its Parts? The Effect of AI Explanations on Complementary Team Performance
Bansal, G., Wu, T., Zhou, J., Fok, R., Nushi, B., Kamar, E., Ribeiro, M. T., & Weld, D. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445717 · Cited by 757
Does the Whole Exceed its Parts? The Effect of AI Explanations on Complementary Team Performance
Bansal, G., Wu, T., Zhou, J., Fok, R., Nushi, B., Kamar, E., Ribeiro, M. T., & Weld, D. S. (2020) · arXiv · doi:10.48550/arxiv.2006.14779 · Cited by 754
Manipulating and Measuring Model Interpretability
Poursabzi-Sangdeh, F., Goldstein, D. G., Hofman, J. M., Wortman Vaughan, J. W., & Wallach, H. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445315 · Cited by 728 · cited 1×
“On the other hand, sometimes humans also show 217:2 Zhuoran Lu, Dakuo Wang, and Ming Yin a degree of over-reliance on AI, as they blindly accept the recommendations of an AI model even when it makes sizable mistakes [8, 14, 74].”
BackgroundManipulating and Measuring Model Interpretability
Poursabzi-Sangdeh, F., Goldstein, D. G., Hofman, J. M., Vaughan, J. W., & Wallach, H. (2018) · arXiv · doi:10.48550/arxiv.1802.07810 · Cited by 711
Understanding the Effect of Accuracy on Trust in Machine Learning Models
Yin, M., Wortman Vaughan, J., & Wallach, H. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300509 · Cited by 527 · cited 1×
“For example, performance indicators and feedback of the AI model, such as its accuracy [50, 112, 113], confidence [75, 116], and the expectation and first impressions of the model competency [44, 67, 69, 92, 92], are shown to significantly impact people’s willingness to rely on the AI model.”
Self-citationProceedings of the ACM Web Conference 2022
(2022) · doi:10.1145/3485447 · Cited by 522
Beyond Accuracy: The Role of Mental Models in Human-AI Team Performance
Bansal, G., Nushi, B., Kamar, E., Lasecki, W. S., Weld, D. S., & Horvitz, E. (2019) · Proceedings of the AAAI Conference on Human Computation and Crowdsourcing · doi:10.1609/hcomp.v7i1.5285 · Cited by 466 · cited 1×
“…performance-related information is absent, it is found that people may utilize other heuristics or cues to determine how to rely on the AI model, including how frequently the AI recommendations align with their own judgments [58] and their mental models of the AI model’s error boundaries [3, 4].”
BackgroundMaking sense of recommendations
Yeomans, M., Shah, A., Mullainathan, S., & Kleinberg, J. (2019) · Journal of Behavioral Decision Making · doi:10.1002/bdm.2118 · Cited by 459 · cited 1×
“Interestingly, both the phenomenon of “algorithm aversion” (i.e., recommendations from humans are used more than those from AI) [21, 24, 111] and “algorithm appreciation” (i.e., recommendations from AI models are used more than those from humans) [55] are observed in different contexts.”
BackgroundThe Role of Explanations on Trust and Reliance in Clinical Decision Support Systems
Bussone, A., Stumpf, S., & O'Sullivan, D. (2015) · 2015 International Conference on Healthcare Informatics · doi:10.1109/ichi.2015.26 · Cited by 441 · cited 2×
“…models for some people [5, 10, 16, 80, 102, 106]—Schaffer et al. [80] showed that showing explanations to people who reported higher familiarity with the decision making tasks led to automation bias, while detailed explanations were also found to lead users to develop over-reliance on AI [10, 72].”
BackgroundThe Principles and Limits of Algorithm-in-the-Loop Decision Making
Green, B., & Chen, Y. (2019) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3359152 · Cited by 392 · cited 1×
“…Carton et al. [11] found that the feature-based explanation does not help people detect online toxic content more accurately when they are assisted by toxic text classifiers, but the same type of explanation was shown to improve people’s accuracy in AI-assisted recidivism risk assessments [31].”
ResultLai, V., & Tan, C. (2019) · Proceedings of the Conference on Fairness, Accountability, and Transparency · doi:10.1145/3287560.3287590 · Cited by 377 · cited 1×
“For example, performance indicators and feedback of the AI model, such as its accuracy [50, 112, 113], confidence [75, 116], and the expectation and first impressions of the model competency [44, 67, 69, 92, 92], are shown to significantly impact people’s willingness to rely on the AI model.”
Human-AI Collaboration in Data Science: Exploring Data Scientists' Perceptions of Automated AI
Wang, D., Weisz, J. D., Muller, M., Ram, P., Geyer, W., Dugan, C., Tausczik, Y., Samulowitz, H., & Gray, A. (2019) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3359313 · Cited by 342
Self-citationWang, X., & Yin, M. (2021) · 26th International Conference on Intelligent User Interfaces · doi:10.1145/3397481.3450650 · Cited by 320 · cited 1×
“…sometimes the provision of certain kinds of explanations may even result in significant over-reliance on AI models for some people [5, 10, 16, 80, 102, 106]—Schaffer et al. [80] showed that showing explanations to people who reported higher familiarity with the decision making tasks led to…”
Self-citationExplaining Decision-Making Algorithms through UI: Strategies to Help Non-Expert Stakeholders
Cheng, H.-F., Wang, R., Zhang, Z., O'Connell, F., Gray, T., Harper, F. M., & Zhu, H. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300789 · Cited by 292 · cited 1×
“Another commonly used intervention is to provide AI explanations along with the decision recommendations, which allow people to probe into the AI model’s rationale before determining whether to rely on its recommendations [5, 12, 13, 23, 49, 53, 99, 106, 116].”
Updates in Human-AI Teams: Understanding and Addressing the Performance/Compatibility Tradeoff
Bansal, G., Nushi, B., Kamar, E., Weld, D. S., Lasecki, W. S., & Horvitz, E. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33012429 · Cited by 284 · cited 1×
“…performance-related information is absent, it is found that people may utilize other heuristics or cues to determine how to rely on the AI model, including how frequently the AI recommendations align with their own judgments [58] and their mental models of the AI model’s error boundaries [3, 4].”
BackgroundKocielnik, R., Amershi, S., & Bennett, P. N. (2019) · Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3290605.3300641 · Cited by 277
Wang, D., Churchill, E., Maes, P., Fan, X., Shneiderman, B., Shi, Y., & Wang, Q. (2020) · Extended Abstracts of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3334480.3381069 · Cited by 271 · cited 1×
“As humans and AI may each possess unique intelligence that is complementary to each other, the human-AI collaboration [97] in this decision-making scenario has the potential to utilize the best of humans and AI and realizes a joint performance beyond what can be achieved by each party alone.”
Self-citationEvaluating Explainable AI: Which Algorithmic Explanations Help Users Predict Model Behavior?
Hase, P., & Bansal, M. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.491 · Cited by 254
Adaptive trust calibration for human-AI collaboration
Okamura, K., & Yamada, S. (2020) · PLOS ONE · doi:10.1371/journal.pone.0229132 · Cited by 239 · cited 1×
“Frameworks have been proposed to monitor people’s trust and reliance on AI and use cognitive cues to prompt them to re-calibrate their trust and reliance when needed [70].”
BackgroundUnderstanding the Role of Human Intuition on Reliance in Human-AI Decision-Making with Explanations
Chen, V., Liao, Q. V., Wortman Vaughan, J., & Bansal, G. (2023) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3610219 · Cited by 220
Appropriate Reliance on AI Advice: Conceptualization and the Effect of Explanations
Schemmer, M., Kuehl, N., Benz, C., Bartos, A., & Satzger, G. (2023) · Proceedings of the 28th International Conference on Intelligent User Interfaces · doi:10.1145/3581641.3584066 · Cited by 208
How do visual explanations foster end users' appropriate trust in machine learning?
Yang, F., Huang, Z., Scholtz, J., & Arendt, D. L. (2020) · Proceedings of the 25th International Conference on Intelligent User Interfaces · doi:10.1145/3377325.3377480 · Cited by 200 · cited 3×
“Another commonly used intervention is to provide AI explanations along with the decision recommendations, which allow people to probe into the AI model’s rationale before determining whether to rely on its recommendations [5, 12, 13, 23, 49, 53, 99, 106, 116].”
Wischnewski, M., Krämer, N., & Müller, E. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581197 · Cited by 146
Ma, S., Lei, Y., Wang, X., Zheng, C., Shi, C., Yin, M., & Ma, X. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581058 · Cited by 143 · cited 1×
“Most recently, researchers have also explored the use of computational approaches to model and predict people’s interactions with AI models [51, 52, 57, 101], which inform the designs of adaptive interfaces to improve people’s appropriate reliance on AI [59].”
Self-citationHuman Reliance on Machine Learning Models When Performance Feedback is Limited: Heuristics and Risks
Lu, Z., & Yin, M. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445562 · Cited by 131
Self-citation"Why is 'Chicago' deceptive?" Towards Building Model-Driven Tutorials for Humans
Lai, V., Liu, H., & Tan, C. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376873 · Cited by 123
I can do better than your AI: expertise and explanations
Schaffer, J., O'Donovan, J., Michaelis, J., Raglin, A., & Höllerer, T. (2019) · Proceedings of the 24th International Conference on Intelligent User Interfaces · doi:10.1145/3301275.3302308 · Cited by 119
Human-Centered Explainable AI (HCXAI): Beyond Opening the Black-Box of AI
Ehsan, U., Wintersberger, P., Liao, Q. V., Watkins, E. A., Manger, C., Daumé III, H., Riener, A., & Riedl, M. O. (2022) · CHI Conference on Human Factors in Computing Systems Extended Abstracts · doi:10.1145/3491101.3503727 · Cited by 117
Nourani, M., King, J., & Ragan, E. (2020) · Proceedings of the AAAI Conference on Human Computation and Crowdsourcing · doi:10.1609/hcomp.v8i1.7469 · Cited by 116
Liu, H., Lai, V., & Tan, C. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3479552 · Cited by 111
Knowing About Knowing: An Illusion of Human Competence Can Hinder Appropriate Reliance on AI Systems
He, G., Kuiper, L., & Gadiraju, U. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581025 · Cited by 107 · cited 1×
“…their interactions with AI models, carefully designed training can help to enhance people’s AI literacy and their overall understanding of the AI model’s behavior [14, 15, 25, 35, 35, 49], which result in a decrease in people’s inappropriate reliance on the AI model during the actual interactions.”
Zhang, S., Yu, J., Xu, X., Yin, C., Lu, Y., Yao, B., Tory, M., Padilla, L. M., Caterino, J., Zhang, P., et al. (2024) · Proceedings of the CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642343 · Cited by 99 · cited 1×
“With its rapid development in recent years, Artificial Intelligence (AI) technology has been integrated into many industries, such as business [39, 41, 79, 96, 98], healthcare [45, 46, 100, 115], education [117], transportation [65], and more.”
BackgroundSelf-citation”Because AI is 100% right and safe”: User Attitudes and Sources of AI Authority in India
Kapania, S., Siy, O., Clapper, G., SP, A. M., & Sambasivan, N. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3517533 · Cited by 90 · cited 1×
“…hand, it is possible that the decision-maker (e.g., the investor) may perceive the AI model to be more competent than their human peers (e.g., AI has the “expert power” or authority) [38, 43]; if so, the presence of second opinions from peers may hardly change how they interact with the AI model.”
BackgroundFok, R., & Weld, D. S. (2024) · AI Magazine · doi:10.1002/aaai.12182 · Cited by 81
Do I trust my machine teammate?: an investigation from perception to decision
Yu, K., Berkovsky, S., Taib, R., Zhou, J., & Chen, F. (2019) · Proceedings of the 24th International Conference on Intelligent User Interfaces · doi:10.1145/3301275.3302277 · Cited by 71 · cited 1×
“For example, performance indicators and feedback of the AI model, such as its accuracy [50, 112, 113], confidence [75, 116], and the expectation and first impressions of the model competency [44, 67, 69, 92, 92], are shown to significantly impact people’s willingness to rely on the AI model.”
MethodSecond Chance for a First Impression? Trust Development in Intelligent System Interaction
Tolmeijer, S., Gadiraju, U., Ghantasala, R., Gupta, A., & Bernstein, A. (2021) · Proceedings of the 29th ACM Conference on User Modeling, Adaptation and Personalization · doi:10.1145/3450613.3456817 · Cited by 67 · cited 1×
“For example, performance indicators and feedback of the AI model, such as its accuracy [50, 112, 113], confidence [75, 116], and the expectation and first impressions of the model competency [44, 67, 69, 92, 92], are shown to significantly impact people’s willingness to rely on the AI model.”
MethodA Slow Algorithm Improves Users' Assessments of the Algorithm's Accuracy
Park, J. S., Barber, R., Kirlik, A., & Karahalios, K. (2019) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3359204 · Cited by 65
Feature-Based Explanations Don't Help People Detect Misclassifications of Online Toxicity
Carton, S., Mei, Q., & Resnick, P. (2020) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v14i1.7282 · Cited by 58 · cited 1×
“For instance, Carton et al. [11] found that the feature-based explanation does not help people detect online toxic content more accurately when they are assisted by toxic text classifiers, but the same type of explanation was shown to improve people’s accuracy in AI-assisted recidivism risk…”
BackgroundYou’d Better Stop! Understanding Human Reliance on Machine Learning Models under Covariate Shift
Chiang, C.-W., & Yin, M. (2021) · 13th ACM Web Science Conference 2021 · doi:10.1145/3447535.3462487 · Cited by 53 · cited 2×
“…their interactions with AI models, carefully designed training can help to enhance people’s AI literacy and their overall understanding of the AI model’s behavior [14, 15, 25, 35, 35, 49], which result in a decrease in people’s inappropriate reliance on the AI model during the actual interactions.”
Self-citationAI-Assisted Human Labeling: Batching for Efficiency without Overreliance
Ashktorab, Z., Desmond, M., Andres, J., Muller, M., Joshi, N. N., Brachman, M., Sharma, A., Brimijoin, K., Pan, Q., Wolf, C. T., et al. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3449163 · Cited by 45
Chiang, C.-W., & Yin, M. (2022) · 27th International Conference on Intelligent User Interfaces · doi:10.1145/3490099.3511121 · Cited by 42
Self-citationAre Visual Explanations Useful? A Case Study in Model-in-the-Loop Prediction
Chu, E., Roy, D., & Andreas, J. (2020) · arXiv · doi:10.48550/arxiv.2007.12248 · Cited by 34 · cited 2×
“…and sometimes the provision of certain kinds of explanations may even result in significant over-reliance on AI models for some people [5, 10, 16, 80, 102, 106]—Schaffer et al. [80] showed that showing explanations to people who reported higher familiarity with the decision making tasks led…”
BackgroundResult
Cited by
Enhancing AI-Assisted Group Decision Making through LLM-Powered Devil's Advocate
Chiang, C.-W., Lu, Z., Li, Z., & Yin, M. (2024) · Proceedings of the 29th International Conference on Intelligent User Interfaces · doi:10.1145/3640543.3645199 · Cited by 142
Self-citationHe, G., Demartini, G., & Gadiraju, U. (2025) · Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706598.3713218 · Cited by 23
Raees, M., Khan, V.-J., Lykourentzou, I., & Papangelis, K. (2026) · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791467 · Cited by 5
“Lu et al. [93] pointed out that active solicitation can also help reduce over-reliance, for instance, by getting a second opinion to form a consensus.”
InfluentialFrom Trust to Appropriate Reliance: Measurement Constructs in Human-AI Decision-Making
Raees, M., & Papangelis, K. (2026) · arXiv · doi:10.48550/arxiv.2604.23896 · Cited by 5
“Lu et al. [49] pointed out that active solicitation can also influence reliance, for instance, by getting a second opinion to form a consensus.”
Zhu, Y., Sui, D., Wang, Z., Hu, X., Gu, L., Qi, Y., Wu, T., Wang, L., Wei, Y., Tang, W., et al. (2026) · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791458 · Cited by 3
Metadata
- Authors
- Lu, Z., Wang, D., & Yin, M.
- Venue
- Proceedings of the ACM on Human-Computer Interaction
- Published
- 2024-04-17
- Volume
- 8
- Issue
- CSCW1
- Pages
- 1-31
- ISSN
- 2573-0142 (ISSN-L)
- OpenAlex
- W4396218105
- Cited by
- 39
- References
- 48
- FWCI
- 18.50
- Type
- journal article
- Language
- en
- Open access
- hybrid
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.
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
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.