Exploring automation bias in human–AI collaboration: a review and implications for explainable AI
Open accessThe DOI matches its Crossref record; 74 of the 74 references it cites were found in a registry or an index; 17 of the 264 sources its claims cite were checked for support.
Abstract
Systematic review of automation bias in high-stakes human-AI collaboration, examining cognitive mechanisms, expertise, trust dynamics, verification demands, and the limits of explainable AI for mitigating overreliance.
Overview
CoChat Summary
## TL;DR
This paper is a PRISMA-based review of automation bias in human-AI collaboration, synthesizing 35 peer-reviewed experimental studies from 2015–April 2025. It concludes that trust, task complexity, AI literacy, expertise, and individual differences shape overreliance, while explainability/transparency alone often do not prevent automation bias and can sometimes worsen it.
## Key findings
- The review included 35 studies, published across 21 journals, with a combined 19,774 participants; all included studies used quantitative, experimental methods.
- Trust and task complexity emerged as the main drivers of automation bias; the paper reports that trust accounted for up to 24.1% of the variance in reliant behavior in one study (Küper et al. 2025).
- Explainability was mixed or limited: CFFs reduced reliance on incorrect AI recommendations compared with simple XAI, but were seen as more cognitively demanding; multiple explanation types did not improve detection of incorrect AI recommendations, and trust calibration feedback did not help.
- The review reports several quantitative examples of bias effects, including Bond et al. (2018) where incorrect automated diagnoses reduced ECG interpretation accuracy from 84.86% to 41.66% for fellows and from 86.38% to 27.43% for non-fellows, and Dratsch et al. (2023) where mammography accuracy dropped from 79.7% to 19.8%, 81.3% to 24.8%, and 82.3% to 45.5% across experience groups when AI predictions were incorrect.
## Methods
- Systematic review conducted using PRISMA 2020; the authors defined five research questions on causal/mediator factors, co-occurring biases, explainability, AI literacy/mental models, and individual differences.
- Searches were run in SCOPUS, ScienceDirect, PubMed, and Google Scholar in May 2025, using peer-reviewed English-language studies from January 2015 to April 2025 that were quantitative, experimental, and focused on automation bias or over-reliance.
- Screening proceeded from title/abstract to full text, with 35 studies included; the two reviewers reported 0.95 Cohen’s kappa agreement.
## Limitations
- The paper states that restricting inclusion to peer-reviewed journal articles from 2015–2025, English-language studies, and experimental designs may have excluded earlier work, non-English studies, preprints, qualitative studies, and mixed-methods research.
- The paper states its search strategy lacked sensitivity and missed some relevant studies due to indexing gaps and terminology variation.
- The paper notes limited evidence on cultural and socio-demographic factors because of the current lack of cross-cultural comparison studies.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
17 Valid · 3 Not valid · 203 Not checked · 48 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“To the best of our knowledge, existing reviews have focused on AB in relation to traditional computer-based aids rather than AI-driven systems (see Goddard etal. 2012; Lyell and Coiera 2017).”
What the sources say
- Not matchedcites Goddard etal. 2012
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Partly supportedcites Lyell and Coiera 2017 sourceExperimental — accuracy not yet measured
The Introduction and Method describe a review of automation bias involving clinical decision support and other automated aids, which supports the characterization of this review's focus. Our copy does not distinguish those aids from AI-driven systems or establish that existing reviews generally excluded AI-driven systems. (LLM, conf=0.82)
hedged- 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.
- Not matched
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Cognitive overload is considered the most significant factor inhibiting users from adequately verifying automation accuracy, thereby increasing AB:”
What the sources say
- Partly supportedcites Lyell and Coiera 2017 sourceExperimental — accuracy not yet measured
The Discussion hypothesizes that cognitive overload can prevent users from adequately verifying automation and lead to automation-bias errors. It does not establish cognitive overload as the most significant factor; the Conclusion says further research is needed to test its impact. The supplied passages do not include page numbers. (LLM, conf=0.88)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Users often fail to critically evaluate automation outputs when they perceive the system to be highly reliable (Lyell and Coiera 2017).”
What the sources say
- Supportedcites Lyell and Coiera 2017 source
The Introduction (page unavailable) says high system accuracy may foster trust and that users with greater trust in automation are less likely to detect its failures. It also describes automation bias as replacing vigilant information seeking with reliance on automated cues, supporting the paraphrase. (LLM, conf=0.90)
plainstates a possible 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.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“the harder the verification complexity, that is the complexity of verifying that the automation is performing correctly, the greater the likelihood of undue trust in the system (Lyell and Coiera 2017).”
What the sources say
- Partly supportedcites Lyell and Coiera 2017 sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) says high verification complexity appears to increase the risk of automation bias, and the Method defines verification complexity as the difficulty of checking whether automation performs correctly. The claim presents that tentative relationship as a definite one; the paper does not establish that likelihood rises consistently as complexity increases. (LLM, conf=0.83)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“In line with these results, Dikmen and Burns (2022) found that providing domain-specific knowledge helped non-expert users to critically evaluate AI outputs, relying less on incorrect AI predictions.”
What the sources say
- Supportedcites Dikmen and Burns 2022 source
Section 6.1 (page unavailable) reports that participants given domain-relevant guidance made fewer investments when the AI underpredicted loan-default risk. It says they may have noticed discrepancies and questioned the AI’s guidance, and concludes that conveying domain knowledge reduced over-reliance on incorrect AI predictions. The finding concerns the peer-to-peer lending task studied. (LLM, conf=0.91)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
132 more claims · 9 with their cited sources checked
Create a free account to see all 271 claims and every source verdict.
- It may state two separate claims as one. 96
- Probably not a substantial claim. 55
- Borderline: one of the tests for a substantial claim is unclear. 2
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?
In the citing paper
“A comparable dynamic may plausibly extend to generative AI used as an information source, where fluent and confident outputs can be accepted uncritically (Romeo and Conti 2026), although this transfer across system classes should be treated with caution.”
From Epistemic Trust in Generative AI as an Information Source: Development and Validation of the Trust in AI-Epistemic Scale (TAI-E) (2026) · cites this paper as Romeo and Conti 2026
In the citing paper
“This configuration does not fit either of two simplified narratives that frequently appear in the surrounding literature: the assumption that analytic engagement necessarily reduces trust (Alon-Barkat and Busuioc 2023; Romeo and Conti 2026;Stanovich et al. 2016), or the assumption that trust primarily reflects uncritical acceptance (Davis 1989;Hoff and Bashir 2015).”
From Epistemic Trust in Generative AI as an Information Source: Development and Validation of the Trust in AI-Epistemic Scale (TAI-E) (2026) · cites this paper as Romeo and Conti 2026
Citations
Cites
The PRISMA 2020 statement: An updated guideline for reporting systematic reviews
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., et al. (2021) · BMJ · doi:10.1136/bmj.n71 · Cited by 108,283
Trust 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 2×
“In the automation context, trust is defined as «the attitude that an (automated) agent will help achieve an individual’s goals in a situation characterized by uncertainty and vulnerability» (Lee and See 2004: p.51).”
The PRISMA 2020 statement: an updated guideline for reporting systematic reviews
Page, M. J., McKenzie, J., Bossuyt, P., Boutron, I., Hoffmann, T., mulrow, C., Shamseer, L., Tetzlaff, J., Akl, E., Brennan, S. E., et al. (2020) · doi:10.31222/osf.io/v7gm2 · Cited by 5,177
Rethlefsen, M. L., Kirtley, S., Waffenschmidt, S., Ayala, A. P., Moher, D., Page, M. J., Koffel, J. B., PRISMA-S Group, Blunt, H., Brigham, T., et al. (2021) · Systematic Reviews · doi:10.1186/s13643-020-01542-z · Cited by 3,714 · cited 2×
“In accordance with PRISMA-S guidelines (Rethlefsen et al. 2021), we also included records identified through personal files.”
Rethlefsen, M. L., Kirtley, S., Waffenschmidt, S., Ayala, A. P., Moher, D., Page, M. J., & Koffel, J. (2019) · doi:10.31219/osf.io/sfc38 · Cited by 3,696
What is AI Literacy? Competencies and Design Considerations
Long, D., & Magerko, B. (2020) · Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3313831.3376727 · Cited by 3,241 · cited 1×
“Key challenges lie in AI literacy, whose lack can limit user’s ability to effectively collaborate with systems (Long and Magerko 2020), and explanation design itself.”
Complacency and Bias in Human Use of Automation: An Attentional Integration
Parasuraman, R., & Manzey, D. H. (2010) · Human Factors: The Journal of the Human Factors and Ergonomics Society · doi:10.1177/0018720810376055 · Cited by 1,710 · cited 1×
“Some researchers emphasize differences in their cognitive and situational antecedents (Wickens et al. 2015), while others highlight substantial overlap, viewing both as manifestations of shared underlying attentional and decision-making mechanisms (Parasuraman and Manzey 2010).”
Automation bias: a systematic review of frequency, effect mediators, and mitigators
Goddard, K., Roudsari, A., & Wyatt, J. C. (2012) · Journal of the American Medical Informatics Association · doi:10.1136/amiajnl-2011-000089 · Cited by 1,162 · cited 7×
“Additionally, accountability, which has been proposed as an intervention to mitigate AB (Goddard et al. 2012), lacks empirical support.”
InfluentialBuçinca, Z., Malaya, M. B., & Gajos, K. Z. (2021) · arXiv · doi:10.48550/arxiv.2102.09692 · Cited by 1,124 · cited 12×
“In an effort to reduce over-reliance, Buçinca et al. (2021) adopted CFFs, a metacognitive debiasing technique aimed at promoting analytical thinking.”
InfluentialZhang, 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 · cited 1×
“Merely making AI explainable or transparent is not enough: explanations must actually help users recognize uncertainty or errors (Zhang et al. 2020).”
Research works printed but not linked: 2
- Hemmer P, Schemmer M, Vössing M, Kühl N (2021) Human–AI complementarity in hybrid intelligence systems: a structured literature review. In: Pacific Asia conference on information systems, vol 78. ISBN 978-1-7336325-7-7
- Romeo G, Conti D (2024) Beyond automation: reshaping Human-artificial intelligence interaction. Sistemi Intelligenti XXVI(3):641–648. https://doi.org/10.1422/115336 · doi:10.1422/115336
Cited by
Toward a science of human–AI teaming for decision making: a complementarity framework
Gonzalez, C., Donahue, K., Goldstein, D. G., Heidari, H., Jalali, M. S., Schelble, B., Singh, A., & Woolley, A. W. (2026) · PNAS Nexus · doi:10.1093/pnasnexus/pgag030 · Cited by 33
“Teams should establish escalation and disagreement protocols (eg confidence thresholds for deferring to humans; concurrence requirements for critical actions) and treat errors as learning opportunities through after-action reviews that incorporate both human and AI system outputs (109).”
Cognitive Risks of AI: Literacy, Trust, and Critical Thinking
Kulal, A. (2025) · Journal of Computer Information Systems · doi:10.1080/08874417.2025.2582050 · Cited by 46
Watts, K. J. (2025) · Education Sciences · doi:10.3390/educsci15101304 · Cited by 18
Cognitive Atrophy Paradox of AI–Human Interaction: From Cognitive Growth and Atrophy to Balance
Kabashkin, I. (2025) · Information · doi:10.3390/info16111009 · Cited by 39
“Human–AI collaboration studies extend these insights, showing that overconfidence in AI may increase even among moderately experienced users, especially when transparency is low, stakes are ambiguous, or monitoring demands are reduced [22].”
Examining human reliance on artificial intelligence in decision making
Pearson, J., Dror, I. E., Jayes, E., Whordley, G.-R., Mason, G., & Nightingale, S. J. (2026) · Scientific Reports · doi:10.1038/s41598-026-34983-y · Cited by 32
“much work has focused on the aviation sector, concerns about automation bias extend to other domains such as healthcare and military, especially with the relatively recent move to embed AI systems within these areas to assist human decision making 12,13 .”
Tres escenarios para la IA en educación: del apoyo responsable a la cocreación
García‐Peñalvo, F. J. (2025) · Education in the Knowledge Society (EKS) · doi:10.14201/eks.32932 · Cited by 19
Wang, C., Yin, Y., & Hu, H. (2026) · Technology in Society · doi:10.1016/j.techsoc.2026.103264 · Cited by 43
Peng, X., Sun, K., Shan, X., & Zhang, J. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1776445 · Cited by 19
“More specifically, when AI-based digital evaluation is uncritically treated as an absolutely objective authority (Romeo and Conti, 2025), or when generative AI directly takes over core decision-making processes (Hou et al., 2025), this kind of substitutive use is more likely to consolidate into…”
From 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
“Romeo et al. [67] reviewed research on “automation bias” influencing humans’ trust and reliance in human-AI collaboration, highlighting the impact of factors such as AI literacy, expertise, and explanations.”
InfluentialEffects of AI explanations on trust and reliance: a study in job shop scheduling
Saßmannshausen, T. M., Burggräf, P., Hassenzahl, M., & Sauer, C. R. (2026) · Ergonomics · doi:10.1080/00140139.2026.2634115 · Cited by 4
When humans stop thinking: tackling the silent threat of AI complacency in service operations
Le, K. B. Q., & Kunz, W. H. (2026) · Journal of service management · doi:10.1108/josm-05-2025-0262 · Cited by 10
Conti, D., Romeo, G., & Nuovo, S. D. (2025) · De Gruyter eBooks · doi:10.1515/9783112206393-002 · Cited by 6
Self-citationDong, A., Li, P., Chen, Y., Gibson, S., Zhao, L., & He, M. (2026) · Sustainability · doi:10.3390/su18115313 · Cited by 9
The double-edged sword of artificial intelligence in invasion biology
Davinack, A. A. (2025) · Biological Invasions · doi:10.1007/s10530-025-03693-3 · Cited by 4
“This amplification matters because outputs are not just faster, they are also often perceived as more authoritative than traditional models (Romeo and Conti 2025), making it more likely that biased results enter decision-making unchecked.”
Generative AI-enabled chatbots for user-centred design: a state-of-the-art review
Zhou, T., Liu, Y., Kumar, M., & Wang, S. (2026) · Journal of Engineering Design · doi:10.1080/09544828.2026.2633488 · Cited by 12
The coalescent architecture of agency: normative directionality as the key to human–AI integration
Torres–Martínez, S. (2026) · AI & Society · doi:10.1007/s00146-026-02979-x · Cited by 4
“As a comprehensive review of automation bias in human–AI collaboration demonstrates, the cognitive mechanisms underlying these errors involve both insufficient attention to contradictory information and failure to search for additional information beyond AI outputs (Romeo and Conti 2024).”
Enhancing graduate students’ evaluative skills through AI-supported collaborative marking
Wu, D., Goh, T., Chen, D., & Yin, H. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1739222 · Cited by 5
“Table 2 presents the coding scheme, developed based on prior work (Heikkinen et al., 2024; Romeo and Conti, 2025; Pavlović, 2025; Wahn et al., 2023; Cheng et al., 2025; Shields et al., 2024; Vo et al., 2024; Onan et al., 2024), to characterize behavioral processes in expert-knowledge–based…”
What is Wrong With Automation Bias?
Jovchevski, P., Buijsman, S., & Neerincx, M. A. (2026) · Philosophy & Technology · doi:10.1007/s13347-026-01090-9 · Cited by 10
“4 Overtrust, and the resulting overreliance, is widely recognized as a primary cause of automation bias across discipline such as ergonomics, legal studies, medicine etc. (Romeo & Conti, 2025 ; Coco, 2023 ; Godard et al. 2012 ; Schemmer et al. 2022 ).”
Investigating How AI Tools Shape Creative Engagement and Self-Efficacy in Design Education
Tianyue, N., Cong, T., Jingru, Z., & Sanmugam, M. (2026) · International Journal of Mobile Human Computer Interaction · doi:10.4018/ijmhci.408631 · Cited by 4
Whose voice counts? The role of large language models in public commenting
Arsenault, A. C., & Kreps, S. (2026) · Big Data & Society · doi:10.1177/20539517261419341 · Cited by 2
Wang, Y., 楊昆岳, & HAN, D. (2026) · Frontiers in Psychology · doi:10.3389/fpsyg.2026.1831770 · Cited by 4
“…research has revealed negative psychological mechanisms including cognitive offloading and the technology convenience paradox (Watts, 2025; Romeo and Conti, 2025): heightened AI awareness may reduce willingness to engage in active thinking, while advanced AI application ability may…”
InfluentialDai, C., Yang, M., & Lee, S. J. (2026) · Systems · doi:10.3390/systems14060639 · Cited by 6
Artificial Intelligence Overload: A Multilevel Taxonomy and the Path Forward
Kazienko, P., Mieleszczenko-Kowszewicz, W., Kocoń, J., et al. (2026) · IEEE Intelligent Systems · doi:10.1109/mis.2026.3666685 · Cited by 2
Cognitive Bias in AI Recommendations: Understanding and Mitigating Human-AI Decision-Making Errors
Aich, P. (2025) · International Conference on Human-Agent Interaction (HAI) · doi:10.1145/3765766.3765887 · Cited by 1
Balázs, L. (2026) · Social Sciences · doi:10.3390/socsci15070433 · Cited by 2
AI in Peer Review: Tool or Threat to Editorial Integrity?
Lu, M. (2025) · Trends in Scholarly Publishing · doi:10.21124/tsp.2025.49.53 · Cited by 2
Morsali, A., Aminia, M., Zarabadi, M. S., Zarabadipour, M., & Mokhlesi, A. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.130150 · Cited by 2
Human-Controllable AI: Meaningful Human Control
Liu, C., & Xu, W. (2025) · Springer eBooks · doi:10.1007/978-981-97-8440-0_58-1 · Cited by 26
Greenwood, J. D., Matthews, M. R., Overgaard, J. D., & Ohde, J. W. (2026) · Mayo Clinic Proceedings Digital Health · doi:10.1016/j.mcpdig.2026.100339 · Cited by 1
Schwarze, M., Mittelstädt, I., Gaidzik, P. W., Kanti, A.-K., & Schiltenwolf, M. (2026) · Die Orthopädie · doi:10.1007/s00132-026-04798-w · Cited by 1
Brunyé, T. T., Petrimoulx, M. V., & Cantelon, J. A. (2026) · Sensors · doi:10.3390/s26072034 · Cited by 3
“That is, people look less, verify less, and accept more when automation appears authoritative [91].”
KI-Algorithmen – intraoperatives hämodynamisches Monitoring von morgen?
Dehbozorgi, M., Zental, N., Larmann, J., & Grüßer, L. (2026) · AINS - Anästhesiologie · Intensivmedizin · Notfallmedizin · Schmerztherapie · doi:10.1055/a-2617-6321 · Cited by 1
Algorithmic Habituation: A Neurocognitive and Systems-Based Framework for Human–AI Co-Adaptation
Mladin, N. C., Rad, D., Coman, D. Ş., Popescu, M. G., Felea, M. I., Marcu, R., & Rad, G. (2026) · Brain Sciences · doi:10.3390/brainsci16050473 · Cited by 3
Pooja, S., Joseph, J., Jose, M., M, A. S., Revati, N., & Joseph, J. (2026) · Frontiers in Developmental Psychology · doi:10.3389/fdpys.2026.1835258 · Cited by 1
Wu, X., & Chen, S. F. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2668041 · Cited by 3
Apologizing artificial intelligence: designing and evaluating effective AI apologies after errors
Turel, O., & Cui, T. (2026) · AI & Society · doi:10.1007/s00146-026-03067-w · Cited by 3
“Last, this study answers the call for research from the X-AI literature to examine specific methods/models of how to interact with the user (De Mulder 2025; Romeo and Conti 2025).”
Avramović, N., Markovic, A. P., Comic, T., Čavoški, S., Zornić, N., & Vujović, V. (2026) · Applied Sciences · doi:10.3390/app16104825 · Cited by 1
Lutz, C., Newlands, G., & Jarrahi, M. H. (2026) · Springer eBooks · doi:10.1007/978-981-95-3658-0_87 · Cited by 1
Banfi, G., & Pedroni, M. (2026) · AI & Society · doi:10.1007/s00146-026-03090-x · Cited by 2
“Concerns identified in the literature have largely centred on technical fallibility, such as the production of inaccurate or biased information (Romeo & Conti 2025), and the risk of excessive cognitive delegation (Formosa et al. 2025).”
From Recommender to Actor: The Normative Boundary When RAG Tools Become Tool-Calling Agents
Rashid, M. T. (2026) · Minds and Machines · doi:10.1007/s11023-026-09782-z · Cited by 5
Human-AI Integration in Cybersecurity: An Industry-Aligned Perspective on Incident Management
Kincl, J., Adam, M., & Pavleska, T. (2026) · International Journal of Information Security · doi:10.1007/s10207-026-01259-5 · Cited by 6
“This human-in-the-loop approach is essential to mitigate risks such as automation bias, over-reliance on AI, and misinterpretation of model outputs [35, 36].”
The missing discipline in AI: a call for behavioural science
Sacher, P. M., Michie, S., Hauser, O., Ferrère, A., Salzer, S., Rodger, A., Borg, J. S., Pogrebna, G., & Murphy, S. A. (2026) · Wellcome Open Research · doi:10.12688/wellcomeopenres.25922.1 · Cited by 1
“15 In mental health contexts, reviews of large language models and conversational systems highlight unresolved clinical risks, including inconsistent or hallucinatory outputs, potential reinforcement of harmful beliefs, and delayed appropriate professional intervention.”
Pratama, R. D., Widiati, U., Basthomi, Y., Khoiri, N. E., Muniroh, S., & Omolu, F. A. (2026) · Journal of Academic Ethics · doi:10.1007/s10805-026-09754-z · Cited by 3
Восприятие агентности в диалоге с ИИ
NOAKK, N. V., & Костина, Т. А. (2026) · Cifrovaâ èkonomika. · doi:10.34706/de-2026-03-07 · Cited by 1
Decoster, R., Toomey, R. J., Davis, M., & RYAN, M. (2026) · Journal of Medical Radiation Sciences · doi:10.1002/jmrs.70109 · Cited by 1
“Previous primary research into how technological change impacts radiographers' reasoning is limited (though not absent; for instance Lundvall, Abrandt-Dahlgren and Wirell [35] did explore the impact of new technology in computed tomography on radiographers' practice in a 2015 qualitative study).”
Pan, Y., Liu, Y., Tang, R., Chang, M., Li, Y., & You, X. (2026) · International Journal of Industrial Ergonomics · doi:10.1016/j.ergon.2026.104010 · Cited by 1
Nanda, P. K. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2697344 · Cited by 1
Batool, S., & Adeel, A. (2026) · Innovation Management Frontiers · doi:10.36923/imf.v1i1.432 · Cited by 1
“AI literacy appears to enable employees to use and evaluate AI effectively, whereas excessive reliance can weaken critical judgment and increase conformity (Liu et al., 2025; Klingbeil et al., 2024; Romeo & Conti, 2026).”
InfluentialKulal, A. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5355424 · Cited by 1
Shimelash, N., Rutunda, S., Menon, V., Emmanuel-Fabula, M., Uwimbabazi, A., Rugege, C., Nshimiyimana, C., Rwema, I. S., Kandekwe, M., Berhe, D. F., et al. (2026) · medRxiv · doi:10.64898/2026.02.16.26346409 · Cited by 1
From Use to Oversight: How Mental Models Influence User Behavior and Output in AI Writing Assistants
Rismani, S., Blodgett, S. L., Liao, Q. V., Olteanu, A., & Moon, A. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791670 · Cited by 4
“[54, 65, 73, 79].”
An examination of the justification for post hoc explanations of artificial intelligence
Hu, D. (2025) · AI & Society · doi:10.1007/s00146-025-02742-8 · Cited by 1
Liability in AI-Enabled Clinical Decision Support: Toward a Tiered Responsibility Model
Rizvi, Z., & Iskandarova, S. (2025) · IEEE Global Conference on Artificial Intelligence and Internet of Things (GCAIoT) · doi:10.1109/gcaiot68269.2025.11275563
“This psychological phenomenon, known as automation bias, can lead to a deskilling of the medical profession and a moral distance between the physician and the patient’s well-being [23].”
Ethische Aspekte des Einsatzes von künstlicher Intelligenz in der Medizin
Heinrichs, B. (2025) · Gefässchirurgie · doi:10.1007/s00772-025-01294-x
The Missing Discipline in AI: A Call for Behavioural SciencePreprint
Sacher, P. M., Michie, S. G., Hauser, O., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6108948
Artificial intelligence driven approaches for responsible urban water management
Ahmed, F. E., & Hilal, N. (2026) · Discover Water · doi:10.1007/s43832-026-00365-8 · Cited by 1
Zhang, J., Yao, Z., Kang, D., Rasiah, R., & Gao, N. (2026) · BMC Psychology · doi:10.1186/s40359-026-04246-6 · Cited by 4
“In highly digitalized work settings, such as Shanghai, reliance on AI may therefore become a routine component of task execution [1, 57].”
Informed Skepticism: Navigating The Code Hallucination Dilemma
Faker, O., & Toker, Ş. (2026) · International Informatics and Software Engineering Conference (IISEC) · doi:10.1109/iisec69317.2026.11418436
“On the other hand, if we have complete distrust, we lose the productivity these systems should give us [16].”
Osei, I., Carie, A., Kanithi, L. P., & Boadu, D. O. (2026) · Frontiers in Human Dynamics · doi:10.3389/fhumd.2026.1795513
Ayasrah, M. N., Al‐Rousan, A. H., Almulla, M. O., & Almulla, A. A. (2026) · Education and Information Technologies · doi:10.1007/s10639-026-13941-1 · Cited by 1
“…(c) calibration of reliance to avoid over-reliance (automation bias) or underuse (algorithm aversion); (d) context-sensitive adaption of instructional decisions; and (e) ethical mediation of consequences for students (Alon-Barkat & Busuioc, 2023; Dietvorst et al. 2015a, b; Romeo and Conti 2025).”
Ouyang, W., Du, H., Han, Y. H., Wang, Z., & He, Y. (2026) · Journal of Eye Movement Research · doi:10.3390/jemr19020038 · Cited by 1
Wischnewski, M., Scharmann, A., Ridder, A., & Krämer, N. C. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790983
“In the context of AI, this effect may be particularly pronounced as users often rely on simple signals to judge complex technologies with which they have limited experience [80].”
The synthetic transformation of the art world: ethical tensions
Calvo, P. (2026) · Humanities and Social Sciences Communications · doi:10.1057/s41599-026-06984-9 · Cited by 1
Alegre, J. C. A., & Mytskevych, J. (2026) · doi:10.2196/preprints.99775
Empowerment over Automation in AI Summarization of Online Reviews
Kordzadeh, N. (2026) · Communications of the ACM · doi:10.1145/3773101 · Cited by 1
“5 Transparency also signals platform integrity.”
Hara, T., Chen, W., & Ota, J. (2026) · CIRP Annals · doi:10.1016/j.cirp.2026.04.094
van Voorst, R. (2026) · AI & Society · doi:10.1007/s00146-026-03081-y
Black-Box Stewardship: AI-Assisted Proxy Voting and Fiduciary DutyPreprint
Souther, M. E. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6732498
From Classroom to Enterprise: Architectural Patterns for LLM Integration
Moriarty, B. (2026) · IEEE International Systems Conference (SysCon) · doi:10.1109/syscon66367.2026.11503623
“…peer-reviewed studies spanning cognitive psychology, human factors, HCI, and neuroscience (2015–2025) finds that trust and task complexity are primary drivers of overreliance, and that neither accountability man-dates nor explanatory interfaces reliably mitigate automation bias across contexts [5].”
Golahi, G., Antonio, F., & Annisa, D. N. (2026) · JMIR Medical Education · doi:10.2196/preprints.100828
Agentic AI and the Algorithm of Good
Gounaris, A., & Kosteletos, G. (2026) · AI · doi:10.3390/ai7090352
“…indicates that healthcare has received the greatest research attention among domains investigating AB (n = 11, 31, 4%), followed by the less studied domains of national security (n = 5), public administration (n = 3), finance (n = 3), human resources (n = 2) and mental healthcare (n = 2) [70].”
InfluentialHuman-Controllable AI: Meaningful Human Control
Liu, C., & Xu, W. (2026) · Springer eBooks · doi:10.1007/978-981-95-3658-0_58
Future Directions—Human-AI Collaboration, Adaptive Explanations, and Regulatory Readiness
Mofrad, F. B., & Yousefzamani, M. (2026) · Springer eBooks · doi:10.1007/978-3-032-15088-2_12
Romeo, G., Conti, D., & Nuovo, S. D. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-6471-1.ch004
Self-citationCirasa, C., Høgsdal, H., & Conti, D. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-6471-1.ch015
Self-citationBetween Civic AI and the Agentic State: The Pathways of Post-Digital Democratic Governance
Pane, S. (2026) · Journal of artificial intelligence for sustainable development. · doi:10.69828/4d4ktj
Sodeify, R., Seyednazari, M. A., Dorosti, A. M., & Nourazarian, A. (2026) · Journal of Materials Science Materials in Medicine · doi:10.1007/s10856-026-07070-x · Cited by 2
“Also, the study documented that trust in AI correlates strongly with automation bias — over-relying on automated recommendations while discounting contradictory clinical judgment [276].”
Sun, X., Wei, S., Pan, T., Wang, Y., Bosch, J. A., Echizen, I., Ali, A. E., & Sugawara, S. (2026) · ACM Conference on User Modeling, Adaptation and Personalization (UMAP) · doi:10.1145/3774935.3806190 · Cited by 1
The Human Experience of Algorithmic HR: A Design-Sensitive, Semi-Structured Review
Mizuki, I., & Khan, Z. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-2303-7.ch002 · Cited by 1
The Irreducible Encounter: Human Skills, Artificial Intelligence, and the Future of Ophthalmic Care
Aslam, T. M. (2026) · Ophthalmology and Therapy · doi:10.1007/s40123-026-01423-z
Mohammad, S. I., Oraini, B. A., Al-Dmour, A., Vasudevan, A., Hussain, Z., Rabby, F., & Bansal, R. (2026) · Knowledge and Process Management · doi:10.1002/kpm.70118 · Cited by 1
Survey evidence demands professional guidance on large language model use in medical physics
Norris, S. A., Kron, T., Douglass, M. J. J., & Badawy, M. K. (2026) · Physica Medica · doi:10.1016/j.ejmp.2026.105835
Trustworthy AI in Business Operations: Explainability, Bias Mitigation, and Compliance in AI Systems
Al-Azazi, S. A. (2026) · World Research of Business Administration Journal · doi:10.56830/wrba03202605
Dell’Oca, S., Rotondi, V., Cutrona, V., Landolfi, G., Bettoni, A., & Canetta, L. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6949565
Sabiri, B., Boulitama, O., Mouchtakir, N., & Karim, S. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-0050-2.ch002
Yong, S., Zhao, M., Martin, A., Flick, C., & Yarosh, S. (2026) · ACM International Conference on Interaction Design and Children (IDC) · doi:10.1145/3773077.3806122
Explainable artificial intelligence in urolithiasis: Applications and future directions
Diamantidis, D. E., Tsakaldimis, G., Kafalis, C., Smyrlis, N. V., Lailisidis, S., Chousein, C., Giannopoulos, S., Georgellis, C., Giannakopoulos, S., & Kalaitzis, C. (2026) · Asian journal of urology · doi:10.1016/j.ajur.2026.01.005
‘Yes, I Am a Slave to AI’: Critical Perspectives of Generative AI in Co-creation and Problem-solving
Oliveira, P. M., Oliveira, P., Amin, I., Abreu, R., & Figueiredo, R. (2026) · Journal of Creative Communications · doi:10.1177/09732586261459522
Chopra, A. (2026) · VINE Journal of Information and Knowledge Management Systems · doi:10.1108/vjikms-07-2025-0288
Khalifah of the Self: Spiritual Leadership as the First Boundary in an AI-Driven World
Herari, N., Santika, A. A., Hassan, Z. B., Anjani, F. F., & Pratama, M. R. K. (2026) · IGI Global eBooks · doi:10.4018/979-8-3373-8192-3.ch008
Lennefer, T. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-29842-3_23
Artificial autonomy and algorithmic paternalism: AI shaping human autonomy and decision-making
Hofmann, B. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1860239 · Cited by 7
“…(Lutz et al., 2025; Gesnot, 2025; Olteanu et al., 2025) is another threat to competency as users may come to accept AI outputs too readily, ceasing to apply critical checks, weakening judgment and making users less able of detecting errors and less creative in deliberation (Romeo and Conti, 2025).”
Human–AI Collaboration in Quality Control Decision-Making
Kayyali, M. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-9725-2.ch007
Waheed, S. M. (2026) · AI & Society · doi:10.1007/s00146-026-03145-z
“…simply achieving augmentation perceptions in the broader sense, as “unengaged augmentation”, though classified as a form of augmentation, is essentially an instance of automation or replacement in reality, covered with a layer of automation complacency (Lebovitz et al. 2022; Romeo & Conti 2025).”
Using Large Language Models to Support Human Decision-Making in SOC Infrastructure Management
Αγαπίου, Σ., Stylianou, S., Peratikou, A., Stavrou, E., & Stavrou, S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-29723-5_8
Co-Rationality Mechanisms in AI Decision Support
Hao, X. (2026) · Journal of Computer Information Systems · doi:10.1080/08874417.2026.2694374
From Words to Decisions: The Cognitive Impact of AI Answer Framing in Higher Education
Leiba, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-30784-2_2
Oh, J. (2026) · The Korean Association of General Education · doi:10.46392/kjge.2026.20.3.81
“그 결 과 사실 왜곡 , 논리적 비약 , 편 향된 전제가 학 습 자의 사고 과정에 식 별 없 이 편 입 될 위 험 이 있다 (Shen et al., 2023; Romeo & Conti, 2026).”
Generative AI-Driven Optimisation Intelligence for Sustainable and Agile Organisation
Jadon, A., & Dixit, A. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-2726-4.ch001
Metadata
- Authors
- Romeo, G., & Conti, D.
- Affiliations
- University of Catania (IT)
- Corresponding author
- Romeo, Giuseppe
- Conti, Daniela
- Venue
- AI & SOCIETY
- Published
- 2025-07-03
- Volume
- 41
- Issue
- 1
- Pages
- 259-278
- ISSN
- 0951-5666 (ISSN-L)
- 1435-5655
- OpenAlex
- W4412013201
- Cited by
- 209
- References
- 60
- Citations by year
- 2026: 194 · 2025: 15
- FWCI
- 228.31
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- AI & Society (publishedVersion, cc-by)
- Topics
- Social Sciences › Human Factors and Ergonomics › Human-Automation Interaction and Safety
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Keywords
- automation bias, human-AI collaboration, explainable AI, trust in AI, AI literacy, explanation complexity, professional expertise, explanation design, user engagement, cognitive mechanisms
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 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-01 (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
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.