Decision Biases Research Articles

Systematic literature review: The implementation of artificial intelligence (IA) sophistication in facilitating the employee recruitment process

This study aims to analyze the application of sophisticated artificial intelligence (AI) in facilitating the employee recruitment process. The research method used is Systematic Literature Review (SLR). Researchers obtained research data from Google Scholar and other research journal article providers over a period of 6 years, namely 2019-2024. 73 journals were obtained after that the data was reduced with several criteria, namely inclusion and exclusion criteria, 10 journals were obtained to be reviewed. The results of this study state that sophisticated AI simplifies the employee recruitment process from the beginning of the recruitment process because it can be more efficient in terms of cost, time, and energy and can reduce the subjectivity of human assessment. However, the final recruitment process still requires the intervention of Human Resources. This is because AI is less able to adapt to culture and language which can cause bias in decision making. AI software that is often used by companies to facilitate the employee recruitment process is Job Agrigator Software and Candidate Assessmant Software.

The Impact of Artificial Intelligence on Legal Practices

The integration of artificial intelligence (AI) into the legal profession is reshaping how legal tasks are performed, enhancing efficiency and streamlining operations. AI applications, such as legal research tools, contract preparation software, and predictive analytics, assist lawyers by automating monotonous tasks, allowing them to focus on more complex aspects of their cases. Key benefits include time savings, improved accuracy in drafting, and reduced bias in decision-making. However, the reliance on AI also raises significant concerns, including the potential erosion of essential legal skills, challenges to privacy and data security, and the diminishing role of human judgment in navigating nuanced legal issues. While AI serves as a valuable tool, it cannot replace the critical thinking, empathy, and intuition that human advocates bring to the profession. Balancing AI's efficiency with the necessity of human insight will be crucial for the future of legal practice.

Educational Roles and Scenarios for Large Language Models: An Ethnographic Research Study of Artificial Intelligence

This paper reviews the theoretical background and potential applications of Large Language Models (LLMs) in educational processes and academic research. Utilizing a novel digital ethnographic approach, we engaged in iterative research with OpenAI’s ChatGPT-4 and Google’s Gemini Ultra—two advanced commercial LLMs. The methodology treated LLMs as research participants, emphasizing the AI-guided perspectives and their envisioned roles in educational settings. Our findings identified the potential LLM roles in educational and research processes and we discussed the AI challenges, which included potential biases in decision-making and AI as a potential source of discrimination and conflict of interest. In addition to practical implications, we used the qualitative research results to advise on the relevant topics for future research.

Trade-offs in AI assistant choice: Do consumers prioritize transparency and sustainability over AI assistant performance?

As artificial intelligence (AI) becomes more integrated into society, concerns have arisen about unintended biases in AI-driven decision-making and the environmental impact of AI technology development. AI assistants such as Siri and Alexa, while helpful, can obscure decision-making and contribute to increased energy use and CO2 emissions. The present study explores whether consumers prioritize transparency and environmental sustainability over performance when choosing AI assistants with conjoint designs. Japanese participants were presented with different AI assistant profiles, varying in performance quality, transparency, cost, and environmental efficiency. The results revealed that Japanese participants prioritized transparency over performance when choosing AI assistants, but they prioritized performance over environmental sustainability. Moreover, future-oriented participants placed more importance on sustainability than those with a present orientation, while participants with an internal locus of control valued transparency more than those with an external locus of control. The findings of this study enhance our understanding of how consumers choose AI options and offer valuable guidance for creating AI systems and communication strategies that work effectively.

Evaluating robo-advisors through behavioral finance: a critical review of technology potential, rationality, and investor expectations

The mini review assesses the value propositions of robo-advisors through the lens of behavioral finance. Despite their promise of data-driven, rational investment strategies, robo-advisors may not fully replicate the personalized service of human financial advisors or eliminate human biases in decision-making. A content analysis of 80 peer-reviewed articles and publications was conducted, focusing on the intersection of financial technology and behavioral finance. Literature was retrieved using The Chicago School University Library's OneSearch and the EBSCO host database, with key terms including “robo-advisor,” “investment behavior,” “risk tolerance,” “financial literacy,” and “affective trust.” The review identifies four key limitations of robo-advisors: (1) their inability to replicate the service-relationship of human advisors; (2) the presence of human bias in supposedly rational algorithms; (3) the inability to minimize market risk; and (4) their limited impact on improving users' financial literacy. Instead, robo-advisors temporarily compensate for a lack of financial knowledge through passive investment strategies. The findings suggest that integrating behavioral finance principles could enhance the predictive power of robo-advisors, though this would introduce additional complexities. The review calls for further research and regulatory measures to ensure that these technologies prioritize investor protection and financial literacy as they continue to evolve.

The Effect of Temporary Group Identity on Adolescent Social Mindfulness Decisions: An Empirical Study Using Team Sports Contexts.

Social mindfulness entails the consideration of the interests of others while respecting their autonomy. Although long-term group identity generates in-group favoritism in social mindfulness decisions, the effect of temporary group identity on social mindfulness remains to be validated. This study examined this effect by recruiting a convenience sample of 74 adolescents from a youth basketball club, who were randomly assigned to either an in-group or out-group decision-making condition. A basketball game scenario was used to establish temporary group identities, referencing the Minimal Group paradigm for grouping while applying the SoMi paradigm in a real-world context. The results showed that adolescents exhibited higher social mindfulness (χ2 = 22.774, df = 1, p < 0.001) and preference-adjusted social mindfulness (χ2 = 6.861, df = 1, p < 0.01) toward in-group compared to out-group members. Conversely, they displayed higher levels of preference-adjusted social hostility toward out-group members (χ2 = 11.291, df = 1, p < 0.01). These findings suggest that even temporary group identities, formed in a sports setting, can lead to intergroup bias in adolescents' social decision-making, promoting goodwill toward the in-group while restricting the autonomy of out-group members.

Sensitive personal health information, public trust, and biases in algorithmic decision-making: public health concerns in digital wellness.

Sensitive personal health information, public trust, and biases in algorithmic decision-making: public health concerns in digital wellness.

Evaluating bias of child maltreatment referrals in encounters for unintentional ingestions

BackgroundUnintentional ingestions commonly result in presentation to the pediatric emergency department (PED). No standardized recommendations exist regarding when to call Child Protective Services (CPS) for unintentional ingestions, which allows for the possibility of biased decision-making. ObjectiveIdentify patient and visit level characteristics associated with CPS referral status in unintentional ingestion encounters. Participants and SettingRetrospective cross-sectional study of children ≤6 years old presenting with unintentional ingestion to a metropolitan PED from 2015 to 2021. MethodsICD-10-CM codes were used to identify unintentional ingestion visits in the electronic health record. CPS referral was determined by manual chart review. Bivariable and multivariable logistic regression models identified patient and visit level characteristics associated with CPS referral. Socioeconomic status (SES) was classified based on zip code, the Child Opportunity Index data, and insurance. ResultsOf 3,836 charts reviewed, 808 met inclusion criteria. CPS referrals were made in 109 (13.5%) visits. Compared to non-Hispanic white patients, non-Hispanic Black patients had 7.37 (95% CI:1.38, 39.26) higher adjusted odds of being referred to CPS after adjusting for triage level, age, and ingestion type. Patients in the Lowest SES group had a 3.01 (95% CI:1.42, 6.40) higher adjusted odds when compared to the Middle/Highest SES group. ConclusionThis study found that CPS referrals in unintentional ingestions occur more often in those identifying as non-Hispanic Black and in those in the Lowest SES group, suggesting clinician bias may impact the decision regarding referrals. To mitigate subjective decision-making, we propose creating standardized, objective CPS referral criteria for unintentional ingestions.

Wireless capsule endoscopy anomaly classification via dynamic multi-task learning

Wireless capsule endoscopy (WCE) provides a painless, non-invasive means for early gastrointestinal disease detection and cancer prevention. However, clinicians must diagnose only about 5% of lesion images from tens of thousands of frames, highlighting the need for computer-assisted diagnostic methods to enhance efficiency and reduce the elevated misdiagnosis rates attributed to visual fatigue. Previous research heavily relied on module design, an effective yet highly coupled method with the baseline and incurring additional computational costs. This paper proposes a dynamic multi-task learning method that combines triplet loss and weighted cross-entropy loss to respectively guide the model in learning compact fine-grained representations and establishing less biased decision boundaries, without incurring additional computational costs. Our method outperforms previous advanced methods on two publicly available datasets, achieving an F1 score of 96.47% on Kvasir-Capsule and an F1 score of 96.75% with an accuracy of 96.72% on CAD-CAP. Visualization of the representations and heatmaps confirms the model’s precision in focusing on the lesion area. The prediction model has been uploaded to https://github.com/xli122/WCE_MTL.

Addressing status quo bias to turn a failing decision system introduction around: the case of DB Schenker

ABSTRACT A new decision system introduction gone wrong can be costly and have further negative consequences. One of the reasons why IT introductions can go wrong is user resistance – often caused by a status quo bias of the users towards the old system. An ironic aspect as decision systems are often especially designed to counter decision biases. This case study presents evidence that targeted countermeasures to status quo bias can help. Based on a review of the available literature, research-based countermeasures were identified and later complemented with practice-based countermeasures from a case study at DB Schenker. With the help of these countermeasures, DB Schenker managed to turn around a new financial reporting system introduction gone wrong.

The structured assessment of violence risk in youth demonstrates measurement invariance between Black and White justice-referred youths.

Identification and implementation of effective methods for reducing racial/ethnic bias and disparities in legal settings are paramount in the United States and other countries. One procedure originally thought to reduce bias in legal decisions is the use of risk assessment instruments, which is now being heavily scrutinized. Measurement invariance, a latent trait technique, is a robust method for assessing one form of bias. Measurement invariance involves determining whether risk items in an instrument appear to be functioning the same between racial or other groups. Thus, the present study examined measurement invariance of the Structured Assessment of Violence Risk in Youth (SAVRY) between non-Latino Black and White youths to examine racial bias. We expected the SAVRY to be invariant (lacking measurement bias) between Black and White youths. The sample included 687 Black and 361 White youths, and the study used a large, multistate data set of SAVRYs conducted by probation officers. We conducted measurement invariance testing in a series of hierarchical steps including testing configural and scalar invariance. The SAVRY demonstrated scalar invariance (equal thresholds for ratings from "low" to "moderate" and "moderate" to "high") for all items except one-community disorganization. The findings lend further credibility to the SAVRY, and the structured professional judgment approach, as a method to assess violence risk and case planning needs among youths involved in the legal system. These findings provide more confidence that significant differences in SAVRY risk level or items between Black and White youths are not based in measurement bias, with the exception of the community disorganization item. Potential fixes discussed include eliminating or deemphasizing this item in final risk level ratings. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

A comprehensive review of behavioral biases in financial decision-making: from classical finance to behavioral finance perspectives

This paper offers a detailed analysis of the evolution of financial decision-making theories, focusing on the shift from classical finance to behavioral finance. Classical finance theories, including the Efficient Market Hypothesis and Modern Portfolio Theory, assume that investors behave rationally and that the market is efficient. However, these theories have faced criticisms highlighting the importance of considering irrational behaviors in financial markets. Behavioral finance addresses this gap by integrating psychological insights into financial decision-making. This study systematically reviews the literature on behavioral biases that affect individual investors, identifying fundamental biases and their impact on investment decisions. The analysis emphasizes the role of cognitive limitations and psychological tendencies in shaping market dynamics, influencing asset pricing, investment strategies, and market returns. The research also notes a shift in focus from market-level outcomes to the behavior of individual investors, with an increase in publications. The paper concludes that understanding investors’ biases is crucial for developing effective risk management strategies and investment recommendations, ultimately leading to improved market performance. The findings underscore the growing importance of behavioral finance in explaining investor behavior and market anomalies, highlighting areas for future research in this evolving field.

Commentary on “AI is Changing the World: For Better or for Worse?”

This commentary explores three fundamental premises surrounding the human-AI partnership. First, a human-AI collaboration is perhaps superior to either working independently, as AI enhances human capabilities but requires oversight to ensure ethical and accurate outcomes. Second, AI's effectiveness is limited by the quality and biases of its training data, which underscores the need for diverse, unbiased datasets. Without proper data, AI could perpetuate flawed or biased decisions, impacting areas such as hiring, healthcare, and empathy-driven interactions. Finally, generative AI is prone to “hallucinations,” where it produces plausible yet incorrect outputs. These errors pose significant risks in high-stakes sectors like healthcare and security. As AI becomes more ingrained in society, these challenges raise ethical concerns around job displacement, loss of human autonomy, and biased decision-making. Here, we also examine the implications of AI hallucinations and model collapse, stressing the importance of continuous human intervention to mitigate AI-driven inaccuracies. Ultimately, a balanced partnership between human judgment and AI's scalability, along with rigorous oversight, is necessary to unlock AI's potential while safeguarding societal values.

Xenophobia in Juror Decision-Making

Abstract With the upsurge of anti-globalizing ideologies and politics, the increasing institutionalization of xenophobia within the legal system has emerged as a pressing concern. Existing law and social science research has underexplored xenophobic bias in the US legal system. This article conceptualizes xenophobic bias as consisting of racism and nationalism. It investigates whether mock jurors reach different verdicts on defendant companies from foreign countries of origin (Japan, France, and China) compared to domestic (US) companies. Using a test simulating a patent lawsuit, the research finds no evidence of general xenophobic bias in juror liability verdict decisions, yet there is a specific bias against the Chinese company when granting damage awards. The similarity-leniency effect that has been established in the previous literature is corroborated in this article. Additionally, political views moderate the effects of the company’s country of origin on juror decisions. This research offers a more nuanced conceptual framework of xenophobic bias in juror decision-making for future law and social science research and informs judicial policies seeking to improve jury instructions and jury selection to reduce xenophobic bias.

Exploring Gender Differences in Investment Decision-Making: The Role of AI in Identifying and Addressing Behavioral Biases

This study investigates how Gender effects investment decision-making and AI's potential to detect and reduce behavioral biases. We utilized a 5-point Likert scale to score demographics, investment decision-making, and behavioral bias identification in 354 people who completed a self-administered online survey. The study examined gender, investment choices, and AI's impact on Behavioral Biases using descriptive correlation and SEM. Men scored slightly higher in Investment Decision-Making than women. (M = 3.7230 vs. M = 3.6784). There is no statistical significance for this difference. (t = 0.550, p = 0.583). SEM indicates gender affects behavioral biases. (β = 0.022, p &lt; 0.05) these biases were significantly good for investment performance (β = 0.821, p &lt; 0.05). The findings reveal how gender and behavioral biases interact, suggesting AI could improve financial decision-making.

Thalamic Roles in Conscious Perception Revealed by Low-Intensity Focused Ultrasound Neuromodulation.

The neural basis of conscious perception remains incompletely understood. While cortical mechanisms of conscious content have been extensively investigated, the role of subcortical structures, including the thalamus, remains less explored. We aim to elucidate the causal contributions of different thalamic regions to conscious perception using transcranial low-intensity focused ultrasound (LIFU) neuromodulation. We hypothesize that modulating different thalamic regions would result in distinct perceptual outcomes. We apply LIFU in human volunteers to investigate region-specific and sonication parameter-dependent effects. We target anterior (transmodal-dominant) and posterior (unimodal-dominant) thalamic regions, further divided into ventral and dorsal regions, while participants perform a near-threshold visual perception task. Task performance is evaluated using Signal Detection Theory metrics. We find that the high duty cycle stimulation of the ventral anterior thalamus enhanced object recognition sensitivity. We also observe a general (i.e., region-independent) effect of LIFU on decision bias (i.e., a tendency toward a particular response) and object categorization accuracy. Specifically, high duty cycle stimulation decreases categorization accuracy, whereas low duty cycle shifts decision bias towards a more conservative stance. In conclusion, our results provide causal insight into the functional organization of the thalamus in shaping human visual experience and highlight the unique role of the transmodal-dominant ventral anterior thalamus.

Behavioral Biases and Judicial Decision-Making in Brazil.

We identify and present Brazil's most common behavioral and heuristic biases in judicial decision-making. Through bibliographic and specific cases, we notice the occurrence of the representativeness heuristic, availability heuristic, anchoring heuristic (anchoring effect), confirmation bias, and affect heuristic bias in Brazilian judicial decisions. We also present the current state of Brazilian legislation and its amendments that aim at impartiality in the production, the assessment of evidence, and the judge's conviction. Finally, we present the suggestions and initiatives that aim to mitigate biases and heuristics in judicial decision-making in Brazil, especially with awareness techniques, the replacement of judges by algorithms, and the review of judicial decisions by collegiate bodies.

Language mediation training and the foreign‑language effect in moral decision‑making*

Abstract Moral cognition is an emerging field investigating the cognitive underpinnings of moral judgement and moral decision-making. Of particular interest for Cognitive Translation and Interpreting Studies (CTIS) is the impact of the use of a foreign language on the cognitive processes associated with moral decision making, what has been labeled “Foreign Language Effect” (FLE). Using a second language has been reported to reduce decision-making biases, doing away with framing effects in risk aversion settings. At the same time, moral dilemmas posed in a second language prompt more utilitarian responses than when presented in the subjects’ native tongues. These results beg the following question for the CTIS scholar: does language mediation training mitigate the FLE? This paper reports on an exploratory study investigating the impact of language mediation training on the FLE. Drawing from Costa et al. (2014), 1st and 4th year, bilingual students in a Translation and Interpreting degree are presented with the trolley dilemma both in their L1, L2, and L3. Initial results seem to indicate that language mediation training might have an impact on the FLE. However, the limitations of the study suggest the need for further exploration of the intersection between language mediation training and the FLE.

How Do Managerial Biases Affect Foreign Market Entry Decisions?

In this practitioner-focused study, we discuss how cognitive biases impair foreign market entry decisions. Focusing on three factors that are central to the evaluation of foreign markets: cost, control, and uncertainty, we explain how decision-making biases undermine their assessment and the ways in which the biases can be mitigated. Specifically, overconfidence bias and the sunk cost fallacy negatively affect the estimation of entry costs which can be mitigated by a strong governance structure and independent reviews; anchoring and confirmation biases undermine the assessment of an appropriate level of control requiring the use of systematic decision frameworks and decision extension tools; and representativeness and availability biases limit the comprehension of environmental uncertainty suggesting mitigation strategies such as the challenging of internal assumptions and devil’s advocacy. By explicating how managerial biases turn into flawed assessments of foreign markets, our study uncovers the mechanisms that explain and potentially remedy cognitive pitfalls during internationalization.

Assessment of Decentralized Decision-Making on Livestock Management in Miombo Woodlands of Tanga and Morogoro regions, Tanzania: Bridging Acts and Practice

Despite legal reforms such as the Local Government (District Authorities) Act of 1982 and the Forest Act of 2002, which decentralize forest management in Tanzania’s villages, the persistent degradation of Miombo woodland forests, primarily due to livestock activities, raises concerns about the effectiveness of these reforms. This study compares the provisions of these acts with on-ground realities through data from 27 Focus Group Discussions (FGDs) and 45 Key Informant Interviews (KIIs). Thematic analysis using NVIVO 12 software identified four key discrepancies: decision-making structures, decision-making processes, gender involvement, and village collaboration. Findings show that villagers often unknowingly delegate legislative power to leaders, decision-making is politicized with minimal stakeholder engagement, gender inclusivity is minimal, and villages manage forests independently rather than collaboratively. These gaps have led to biased decisions, conflicts among user groups, forest encroachment, and the neglect of women's needs, exacerbating forest degradation. To address these issues, the study recommends capacity building through leadership and technical training for village councils, community education on legal rights, and improved transparency via public forums and accessible reporting. These initiatives aim to empower local communities and foster sustainable management of Miombo woodland forests