Pub Date : 2025-12-01DOI: 10.1016/j.chbah.2025.100241
Yiwen Jin , Lies Sercu , Feng Guo
As large language models (LLMs) such as ChatGPT are increasingly used across cultures and languages, concerns have arisen about their ability to respond in culturally sensitive ways. This study evaluated the intercultural sensitivity of GPT-3.5 and GPT-4 using the Intercultural Sensitivity Scale (ISS) translated into eight languages. Each model completed ten randomized iterations of the 24-item ISS per language, and the results were analyzed using descriptive statistics and three-way ANOVA. GPT-4 achieved significantly higher intercultural sensitivity scores than GPT-3.5 across all dimensions, with “respect for cultural differences” scoring highest and “interaction confidence” lowest. Significant interactions were found between model version and language, and between model version and ISS dimensions, indicating that GPT-4's improvements vary by linguistic context. Nonetheless, the interaction between language and dimensions did not yield significant results. Future research should focus on increasing the amount of training data for the less spoken languages, as well as adding rich emotional and cultural background data to improve the model's understanding of cultural norms and nuances.
{"title":"Assessing intercultural sensitivity in large language models: A comparative study of GPT-3.5 and GPT-4 across eight languages","authors":"Yiwen Jin , Lies Sercu , Feng Guo","doi":"10.1016/j.chbah.2025.100241","DOIUrl":"10.1016/j.chbah.2025.100241","url":null,"abstract":"<div><div>As large language models (LLMs) such as ChatGPT are increasingly used across cultures and languages, concerns have arisen about their ability to respond in culturally sensitive ways. This study evaluated the intercultural sensitivity of GPT-3.5 and GPT-4 using the Intercultural Sensitivity Scale (ISS) translated into eight languages. Each model completed ten randomized iterations of the 24-item ISS per language, and the results were analyzed using descriptive statistics and three-way ANOVA. GPT-4 achieved significantly higher intercultural sensitivity scores than GPT-3.5 across all dimensions, with “respect for cultural differences” scoring highest and “interaction confidence” lowest. Significant interactions were found between model version and language, and between model version and ISS dimensions, indicating that GPT-4's improvements vary by linguistic context. Nonetheless, the interaction between language and dimensions did not yield significant results. Future research should focus on increasing the amount of training data for the less spoken languages, as well as adding rich emotional and cultural background data to improve the model's understanding of cultural norms and nuances.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100241"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145618013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-01DOI: 10.1016/j.chbah.2025.100242
Rong Chang
{"title":"Aesthetic Integrity Index (AII) for human–AI hybrid epistemology: Reconfiguring the Beholder’s Share through Xie He’s Six Principles","authors":"Rong Chang","doi":"10.1016/j.chbah.2025.100242","DOIUrl":"10.1016/j.chbah.2025.100242","url":null,"abstract":"","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100242"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145618011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-19DOI: 10.1016/j.chbah.2025.100236
Kevin Riehl, Anastasios Kouvelas, Michail A. Makridis
Advancements in computer science, artificial intelligence, and control systems have catalyzed the emergence of cybernetic societies, where algorithms play a pivotal role in decision-making processes shaping nearly every aspect of human life. Automated decision-making for resource allocation has expanded into industry, government processes, critical infrastructures, and even determines the very fabric of social interactions and communication. While these systems promise greater efficiency and reduced corruption, misspecified cybernetic mechanisms harbor the threat for reinforcing inequities, discrimination, and even dystopian or totalitarian structures. Fairness thus becomes a crucial component in the design of cybernetic systems, to promote cooperation between selfish individuals, to achieve better outcomes at the system level, to confront public resistance, to gain trust and acceptance for rules and institutions, to perforate self-reinforcing cycles of poverty through social mobility, to incentivize motivation, contribution and satisfaction of people through inclusion, to increase social-cohesion in groups, and ultimately to improve life quality. Quantitative descriptions of fairness are crucial to reflect equity into algorithms, but only few works in the fairness literature offer such measures; the existing quantitative measures in the literature are either too application-specific, suffer from undesirable characteristics, or are not ideology-agnostic. This study proposes a quantitative, transactional, and distributive fairness framework based on an interdisciplinary foundation that supports the systematic design of socially-feasible decision-making systems. Moreover, it emphasizes the importance of fairness and transparency when designing algorithms for equitable, cybernetic societies, and establishes a connection between fairness literature and resource allocating systems.
{"title":"Quantitative fairness—A framework for the design of equitable cybernetic societies","authors":"Kevin Riehl, Anastasios Kouvelas, Michail A. Makridis","doi":"10.1016/j.chbah.2025.100236","DOIUrl":"10.1016/j.chbah.2025.100236","url":null,"abstract":"<div><div>Advancements in computer science, artificial intelligence, and control systems have catalyzed the emergence of cybernetic societies, where algorithms play a pivotal role in decision-making processes shaping nearly every aspect of human life. Automated decision-making for resource allocation has expanded into industry, government processes, critical infrastructures, and even determines the very fabric of social interactions and communication. While these systems promise greater efficiency and reduced corruption, misspecified cybernetic mechanisms harbor the threat for reinforcing inequities, discrimination, and even dystopian or totalitarian structures. Fairness thus becomes a crucial component in the design of cybernetic systems, to promote cooperation between selfish individuals, to achieve better outcomes at the system level, to confront public resistance, to gain trust and acceptance for rules and institutions, to perforate self-reinforcing cycles of poverty through social mobility, to incentivize motivation, contribution and satisfaction of people through inclusion, to increase social-cohesion in groups, and ultimately to improve life quality. Quantitative descriptions of fairness are crucial to reflect equity into algorithms, but only few works in the fairness literature offer such measures; the existing quantitative measures in the literature are either too application-specific, suffer from undesirable characteristics, or are not ideology-agnostic. This study proposes a quantitative, transactional, and distributive fairness framework based on an interdisciplinary foundation that supports the systematic design of socially-feasible decision-making systems. Moreover, it emphasizes the importance of fairness and transparency when designing algorithms for equitable, cybernetic societies, and establishes a connection between fairness literature and resource allocating systems.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100236"},"PeriodicalIF":0.0,"publicationDate":"2025-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579165","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-19DOI: 10.1016/j.chbah.2025.100238
Ying Qin, Wanhui Zhou, Bu Zhong
Understanding user responses to AI versus human errors is crucial, as they shape trust, acceptance, and interaction outcomes. This study investigates the emotional dynamics of human-AI interactions by examining how agent identity (human vs. AI) and error severity (low vs. high) influence negative emotional reactions. Using a 2 × 2 factorial design (N = 250), the findings reveal that human agents consistently elicit stronger negative emotions than AI agents, regardless of error severity. Moreover, perceived experience moderates this relationship under specific conditions: individuals who view AI less experienced than humans exhibit stronger negative emotions toward human errors, while this effect diminishes when AI is perceived as having higher experience. However, perceived agency does not significantly influence emotional responses. These findings highlight the critical role of agent identity and perceived experience in shaping emotional reactions to errors, adding insights into the dynamics of human-AI interactions. This research shows that developing effective AI systems needs to manage user emotional responses and trust, in which perceived experience and competency play pivotal roles in adoption. The findings can guide the design of AI systems that adjust user expectations and emotional responses in accordance with the AI's perceived level of experience.
{"title":"Why human mistakes hurt more? Emotional responses in human-AI errors","authors":"Ying Qin, Wanhui Zhou, Bu Zhong","doi":"10.1016/j.chbah.2025.100238","DOIUrl":"10.1016/j.chbah.2025.100238","url":null,"abstract":"<div><div>Understanding user responses to AI versus human errors is crucial, as they shape trust, acceptance, and interaction outcomes. This study investigates the emotional dynamics of human-AI interactions by examining how agent identity (human vs. AI) and error severity (low vs. high) influence negative emotional reactions. Using a 2 × 2 factorial design (<em>N</em> = 250), the findings reveal that human agents consistently elicit stronger negative emotions than AI agents, regardless of error severity. Moreover, perceived experience moderates this relationship under specific conditions: individuals who view AI less experienced than humans exhibit stronger negative emotions toward human errors, while this effect diminishes when AI is perceived as having higher experience. However, perceived agency does not significantly influence emotional responses. These findings highlight the critical role of agent identity and perceived experience in shaping emotional reactions to errors, adding insights into the dynamics of human-AI interactions. This research shows that developing effective AI systems needs to manage user emotional responses and trust, in which perceived experience and competency play pivotal roles in adoption. The findings can guide the design of AI systems that adjust user expectations and emotional responses in accordance with the AI's perceived level of experience.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100238"},"PeriodicalIF":0.0,"publicationDate":"2025-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-19DOI: 10.1016/j.chbah.2025.100240
Eun Go , Taeyoung Kim
Despite the widespread use of large language model (LLM)-based chatbots, little is known about what specific gratifications users obtain from the unique affordances of these systems and how these affordance-driven gratifications shape user evaluations. To address this gap, the present study maps the gratification structure of LLM chatbot use and examines whether users’ primary purpose of chatbot use (information-, conversation-, or task-oriented) influences the gratifications they derive. A survey of 249 LLM chatbot users revealed nine distinct gratifications aligned with four affordance types: modality, agency, interactivity, and navigability. Purpose of use meaningfully shaped which gratifications were most salient. For example, conversational use heightened Immersive Realism and Fun, whereas information- and task-oriented use elevated Adaptive Responsiveness. In turn, these affordance-driven gratifications predicted key outcomes, including perceived expertise, perceived friendliness, satisfaction, attitudes, and behavioral intentions to continued use. Across outcomes, Adaptive Responsiveness consistently emerged as the strongest predictor, underscoring the pivotal role of contingent, high-quality dialogue in LLM-based human–AI interaction. These findings extend uses and gratifications theory and offer design implications for developing more engaging, responsive, and purpose-tailored chatbot experiences.
{"title":"Mapping user gratifications in the age of LLM-based chatbots: An affordance perspective","authors":"Eun Go , Taeyoung Kim","doi":"10.1016/j.chbah.2025.100240","DOIUrl":"10.1016/j.chbah.2025.100240","url":null,"abstract":"<div><div>Despite the widespread use of large language model (LLM)-based chatbots, little is known about what specific gratifications users obtain from the unique affordances of these systems and how these affordance-driven gratifications shape user evaluations. To address this gap, the present study maps the gratification structure of LLM chatbot use and examines whether users’ primary purpose of chatbot use (information-, conversation-, or task-oriented) influences the gratifications they derive. A survey of 249 LLM chatbot users revealed nine distinct gratifications aligned with four affordance types: modality, agency, interactivity, and navigability. Purpose of use meaningfully shaped which gratifications were most salient. For example, conversational use heightened <em>Immersive Realism</em> and <em>Fun</em>, whereas information- and task-oriented use elevated <em>Adaptive Responsiveness</em>. In turn, these affordance-driven gratifications predicted key outcomes, including perceived expertise, perceived friendliness, satisfaction, attitudes, and behavioral intentions to continued use. Across outcomes, <em>Adaptive Responsiveness</em> consistently emerged as the strongest predictor, underscoring the pivotal role of contingent, high-quality dialogue in LLM-based human–AI interaction. These findings extend uses and gratifications theory and offer design implications for developing more engaging, responsive, and purpose-tailored chatbot experiences.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"7 ","pages":"Article 100240"},"PeriodicalIF":0.0,"publicationDate":"2025-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145698045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-17DOI: 10.1016/j.chbah.2025.100237
Ruoyu Niu, Mengzhu Huang, Rixin Tang
Virtual gaming worlds support rich social interaction in which players use avatars to collaborate, compete, and communicate across distance. Motivated by the increasing reliance on mediated social contact, this research examined whether virtual shared space and avatar properties shape personal space regulation in ways that parallel face to face encounters. Three experiments tested how virtual shared space, avatar agency, and avatar anthropomorphism influence interpersonal distance. Across studies, virtual comfort distance and psychological distance were used as complementary indicators of changes in personal space, and physical comfort distance was additionally assessed in a subset of conditions with a physically present human partner. Experiment 1 showed that, when interacting with a human driven partner in the laboratory, occupying a shared virtual space reliably reduced comfort distance and increased psychological closeness compared with interacting in separate virtual spaces, even after controlling for physical shared space. Experiment 2 replicated the virtual shared space effect with computer driven partners in an online Virtual gaming world setting, indicating that reduced interpersonal distance does not depend on human agency alone. Experiment 3 revealed that anthropomorphic avatars increased comfort toward computer driven partners, whereas avatar form had little impact when the partner was known to be human. Together, the findings indicate that virtual shared space, perceived agency, and avatar appearance jointly shape personal space regulation in digital environments and offer actionable guidance for designing avatars and virtual spaces that foster approach oriented, prosocial interaction.
{"title":"Avatar or human, who is experiencing it? Impact of social interaction in virtual gaming worlds on personal space","authors":"Ruoyu Niu, Mengzhu Huang, Rixin Tang","doi":"10.1016/j.chbah.2025.100237","DOIUrl":"10.1016/j.chbah.2025.100237","url":null,"abstract":"<div><div>Virtual gaming worlds support rich social interaction in which players use avatars to collaborate, compete, and communicate across distance. Motivated by the increasing reliance on mediated social contact, this research examined whether virtual shared space and avatar properties shape personal space regulation in ways that parallel face to face encounters. Three experiments tested how virtual shared space, avatar agency, and avatar anthropomorphism influence interpersonal distance. Across studies, virtual comfort distance and psychological distance were used as complementary indicators of changes in personal space, and physical comfort distance was additionally assessed in a subset of conditions with a physically present human partner. Experiment 1 showed that, when interacting with a human driven partner in the laboratory, occupying a shared virtual space reliably reduced comfort distance and increased psychological closeness compared with interacting in separate virtual spaces, even after controlling for physical shared space. Experiment 2 replicated the virtual shared space effect with computer driven partners in an online Virtual gaming world setting, indicating that reduced interpersonal distance does not depend on human agency alone. Experiment 3 revealed that anthropomorphic avatars increased comfort toward computer driven partners, whereas avatar form had little impact when the partner was known to be human. Together, the findings indicate that virtual shared space, perceived agency, and avatar appearance jointly shape personal space regulation in digital environments and offer actionable guidance for designing avatars and virtual spaces that foster approach oriented, prosocial interaction.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100237"},"PeriodicalIF":0.0,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-17DOI: 10.1016/j.chbah.2025.100230
Kyana H.J. van Eijndhoven , Ethel Pruss , Pieter Spronck
A growing body of research has focused on examining the role of nonverbal mimicry, the spontaneous imitation of others’ physical behavior during social interactions, in human-virtual human interaction. The increasing deployment of virtual humans, and growing advancements in technology vital to virtual human development, emphasize the necessity to review studies incorporating such state-of-the-art technologies. To this end, we conducted a scoping review of empirical work studying nonverbal mimicry in human-virtual human interaction. This review focused on outlining (1) the contexts in which such interactions occurred, (2) implementations of nonverbal mimicry, (3) individual and situational factors that can lead one to mimic more (facilitators) or less (inhibitors), and (4) individual and social consequences. By creating this comprehensive outline, we were able to capture the current state of nonverbal mimicry research, and identify methodological, evidence, and empirical research gaps, that may serve as future guidelines to drive the field of virtual human research forward.
{"title":"A scoping review of nonverbal mimicry in human-virtual human interaction","authors":"Kyana H.J. van Eijndhoven , Ethel Pruss , Pieter Spronck","doi":"10.1016/j.chbah.2025.100230","DOIUrl":"10.1016/j.chbah.2025.100230","url":null,"abstract":"<div><div>A growing body of research has focused on examining the role of nonverbal mimicry, the spontaneous imitation of others’ physical behavior during social interactions, in human-virtual human interaction. The increasing deployment of virtual humans, and growing advancements in technology vital to virtual human development, emphasize the necessity to review studies incorporating such state-of-the-art technologies. To this end, we conducted a scoping review of empirical work studying nonverbal mimicry in human-virtual human interaction. This review focused on outlining (1) the contexts in which such interactions occurred, (2) implementations of nonverbal mimicry, (3) individual and situational factors that can lead one to mimic more (facilitators) or less (inhibitors), and (4) individual and social consequences. By creating this comprehensive outline, we were able to capture the current state of nonverbal mimicry research, and identify methodological, evidence, and empirical research gaps, that may serve as future guidelines to drive the field of virtual human research forward.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100230"},"PeriodicalIF":0.0,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-17DOI: 10.1016/j.chbah.2025.100239
Emily Lochner , René Schmoll , Stephan Kaiser
As artificial intelligence (AI) becomes increasingly integrated into organizational leadership, it is critical to understand how algorithmic decision-making affects employee well-being. This study investigates how varying levels of AI involvement in leadership – ranging from fully human to hybrid (human-AI collaboration) to fully automated – influence employees' emotional responses at work. It also examines whether the emotional impact of leader type depends on the outcome of a managerial decision (positive vs. negative). To investigate these questions, we conducted a vignette-based online experiment using a 3x2 between-subjects design. Participants (N = 153 workers) were randomly assigned to one of six short, standardized leadership scenarios that varied by leader type (human, hybrid, or AI) and decision outcome (positive or negative). The vignettes described a realistic workplace situation in which a leader communicates a decision about a project's continuation. Subsequently, emotional responses were measured using validated affective scales.
The results showed that higher AI involvement led to lower positive affect, particularly following favorable decisions, while negative affect remained largely unaffected. These results suggest that, while AI leadership is not emotionally harmful, it also fails to generate positive engagement. Positive affect was strongest when positive decisions were delivered by a human leader and weakest when delivered by an AI.
These findings contribute to leadership and human-AI interaction research by highlighting an emotional asymmetry in AI-led leadership. Practically speaking, these results imply that while AI offers efficiency, it lacks the interpersonal resonance necessary for emotionally meaningful interactions. Therefore, organizations should consider maintaining human involvement in contexts where recognition, trust, or relational sensitivity are important.
{"title":"Less human, less positive? How AI involvement in leadership shapes employees’ affective well-being across different supervisor decisions","authors":"Emily Lochner , René Schmoll , Stephan Kaiser","doi":"10.1016/j.chbah.2025.100239","DOIUrl":"10.1016/j.chbah.2025.100239","url":null,"abstract":"<div><div>As artificial intelligence (AI) becomes increasingly integrated into organizational leadership, it is critical to understand how algorithmic decision-making affects employee well-being. This study investigates how varying levels of AI involvement in leadership – ranging from fully human to hybrid (human-AI collaboration) to fully automated – influence employees' emotional responses at work. It also examines whether the emotional impact of leader type depends on the outcome of a managerial decision (positive vs. negative). To investigate these questions, we conducted a vignette-based online experiment using a 3x2 between-subjects design. Participants (N = 153 workers) were randomly assigned to one of six short, standardized leadership scenarios that varied by leader type (human, hybrid, or AI) and decision outcome (positive or negative). The vignettes described a realistic workplace situation in which a leader communicates a decision about a project's continuation. Subsequently, emotional responses were measured using validated affective scales.</div><div>The results showed that higher AI involvement led to lower positive affect, particularly following favorable decisions, while negative affect remained largely unaffected. These results suggest that, while AI leadership is not emotionally harmful, it also fails to generate positive engagement. Positive affect was strongest when positive decisions were delivered by a human leader and weakest when delivered by an AI.</div><div>These findings contribute to leadership and human-AI interaction research by highlighting an emotional asymmetry in AI-led leadership. Practically speaking, these results imply that while AI offers efficiency, it lacks the interpersonal resonance necessary for emotionally meaningful interactions. Therefore, organizations should consider maintaining human involvement in contexts where recognition, trust, or relational sensitivity are important.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100239"},"PeriodicalIF":0.0,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579170","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-13DOI: 10.1016/j.chbah.2025.100233
Tengfei Yu , Siyu Pan , Caoyun Fan , Siyang Luo , Yaohui Jin , Binglei Zhao
Empathy, a key component of human-social interaction, has become a core con-cern in human-computer interaction. This study examines whether current large language models (LLMs) can exhibit empathy in both cognitive and affective dimensions as humans. In our study, we used the standardized questionnaire to assess LLMs empathy ability and a novel paradigm was developed for LLMs eval-uation. Four main experiments were reported on LLMs empathy abilities using the Interpersonal Reactivity Index (IRI) and the Basic Empathy Scale (BES) on GPT-4 and Llama3 respectively. Two levels of evaluations were conducted to investigate whether the structural validity of the questionnaire in LLMs was aligned with humans and to compare the LLMs' empathy abilities with humans. We found GPT-4 show identical empathy dimension structure with humans while exhibiting significantly lower empathy abilities as compared to humans. Moreover, systemati-cal difference empathy ability was evident in Llama3 showing its failure to exhibit the same empathy dimensions as humans. All these findings indicate that though GPT-4 kept the same structure of human empathy (cognitive and affective), the current LLMs can not simulate empathy as we humans as indexed by the response to the questionnaire. This highlights the urgent requirements for further improving LLMs’ empathy abilities for more user-friendly human-LLMs interactions. In addition, the pipeline to generate diverse LLMs-simulated participants was also discussed.
{"title":"Can large language models exhibit cognitive and affective empathy as humans?","authors":"Tengfei Yu , Siyu Pan , Caoyun Fan , Siyang Luo , Yaohui Jin , Binglei Zhao","doi":"10.1016/j.chbah.2025.100233","DOIUrl":"10.1016/j.chbah.2025.100233","url":null,"abstract":"<div><div>Empathy, a key component of human-social interaction, has become a core con-cern in human-computer interaction. This study examines whether current large language models (LLMs) can exhibit empathy in both cognitive and affective dimensions as humans. In our study, we used the standardized questionnaire to assess LLMs empathy ability and a novel paradigm was developed for LLMs eval-uation. Four main experiments were reported on LLMs empathy abilities using the Interpersonal Reactivity Index (IRI) and the Basic Empathy Scale (BES) on GPT-4 and Llama3 respectively. Two levels of evaluations were conducted to investigate whether the structural validity of the questionnaire in LLMs was aligned with humans and to compare the LLMs' empathy abilities with humans. We found GPT-4 show identical empathy dimension structure with humans while exhibiting significantly lower empathy abilities as compared to humans. Moreover, systemati-cal difference empathy ability was evident in Llama3 showing its failure to exhibit the same empathy dimensions as humans. All these findings indicate that though GPT-4 kept the same structure of human empathy (cognitive and affective), the current LLMs can not simulate empathy as we humans as indexed by the response to the questionnaire. This highlights the urgent requirements for further improving LLMs’ empathy abilities for more user-friendly human-LLMs interactions. In addition, the pipeline to generate diverse LLMs-simulated participants was also discussed.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100233"},"PeriodicalIF":0.0,"publicationDate":"2025-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579166","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-01DOI: 10.1016/j.chbah.2025.100225
Vaclav Moravec , Beata Gavurova , Martin Rigelsky
The main goal of the study was to examine and to evaluate the relationships between public attitudes towards AI avatars, their selected socio-demographic characteristics, fields of their media consumption, as well as ideological attitudes in order to reveal other adoption perspectives of AI avatars, which are not yet explored, and their strong economic and social potential in the metaverse. The data collection was carried out on a sample of 1250 respondents aged 18 and over in the period from 2 April 2025 to 9 April 2025. The research used an AI avatar experimentally developed by the start-up company The MAMA AI.
The outcomes of the descriptive analysis confirmed the fact that the AI news avatar Pavel was perceived rather neutrally to slightly positively, but as impersonal, with respondents demonstrating a low willingness to accept him as a guide across the media. The respondents also evaluated the use of AI assistants most favorably in the technical-service fields, but significantly more negatively in the sensitive domains such as psychology or politics. The differences between these groups were most noticeable in the perception of the AI avatar as more or less human and intimate, especially between men and women. Contrariwise, media habits played a much larger role. The study confirmed the importance of investigating specific adoption factors related to media consumption, media habits, and ideological attitudes along with the socio-demographic factors and thus, allowed us to understand the new adoption potential of AI avatars and the possibilities of its expansion.
该研究的主要目标是检查和评估公众对人工智能化身的态度、他们所选择的社会人口特征、媒体消费领域以及意识形态态度之间的关系，以揭示尚未探索的人工智能化身的其他采用观点，以及它们在虚拟世界中强大的经济和社会潜力。在2025年4月2日至2025年4月9日期间，对1250名18岁及以上的受访者进行了数据收集。这项研究使用了初创公司The MAMA AI实验性开发的人工智能化身。描述性分析的结果证实了这样一个事实，即人们对人工智能新闻化身帕维尔的看法相当中性或略微积极，但没有人情感，受访者表示不太愿意接受他作为整个媒体的向导。受访者还对人工智能助手在技术服务领域的使用进行了最有利的评估，但在心理学或政治等敏感领域则明显更为负面。这些群体之间的差异最明显的是对人工智能化身的看法，尤其是在男性和女性之间。相反，媒体习惯发挥了更大的作用。该研究证实了调查与媒体消费、媒体习惯、意识形态态度以及社会人口因素相关的具体采用因素的重要性，从而使我们能够了解人工智能化身的新采用潜力及其扩展的可能性。
{"title":"“What is the latest news, Avatar Pavel?” - AI assistants in transformation processes of metaverse","authors":"Vaclav Moravec , Beata Gavurova , Martin Rigelsky","doi":"10.1016/j.chbah.2025.100225","DOIUrl":"10.1016/j.chbah.2025.100225","url":null,"abstract":"<div><div>The main goal of the study was to examine and to evaluate the relationships between public attitudes towards AI avatars, their selected socio-demographic characteristics, fields of their media consumption, as well as ideological attitudes in order to reveal other adoption perspectives of AI avatars, which are not yet explored, and their strong economic and social potential in the metaverse. The data collection was carried out on a sample of 1250 respondents aged 18 and over in the period from 2 April 2025 to 9 April 2025. The research used an AI avatar experimentally developed by the start-up company The MAMA AI.</div><div>The outcomes of the descriptive analysis confirmed the fact that the AI news avatar Pavel was perceived rather neutrally to slightly positively, but as impersonal, with respondents demonstrating a low willingness to accept him as a guide across the media. The respondents also evaluated the use of AI assistants most favorably in the technical-service fields, but significantly more negatively in the sensitive domains such as psychology or politics. The differences between these groups were most noticeable in the perception of the AI avatar as more or less human and intimate, especially between men and women. Contrariwise, media habits played a much larger role. The study confirmed the importance of investigating specific adoption factors related to media consumption, media habits, and ideological attitudes along with the socio-demographic factors and thus, allowed us to understand the new adoption potential of AI avatars and the possibilities of its expansion.</div></div>","PeriodicalId":100324,"journal":{"name":"Computers in Human Behavior: Artificial Humans","volume":"6 ","pages":"Article 100225"},"PeriodicalIF":0.0,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145465779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号