Human Behavior and Emerging Technologies最新文献

Deepfakes refer to a wide range of computer-generated synthetic media, in which a person’s appearance or likeness is altered to resemble that of another. This systematic review is aimed at providing an overview of the existing research into people’s ability to detect deepfakes. Five databases (IEEE, ProQuest, PubMed, Web of Science, and Scopus) were searched up to December 2023. Studies were included if they (1) were an original study; (2) were reported in English; (3) examined people’s detection of deepfakes; (4) examined the influence of an intervention, strategy, or variable on deepfake detection; and (5) reported relevant data needed to evaluate detection accuracy. Forty independent studies from 30 unique records were included in the review. Results were narratively summarized, with key findings organized based on the review’s research questions. Studies used different performance measures, making it difficult to compare results across the literature. Detection accuracy varies widely, with some studies showing humans outperforming AI models and others indicating the opposite. Detection performance is also influenced by person-level (e.g., cognitive ability, analytical thinking) and stimuli-level factors (e.g., quality of deepfake, familiarity with the subject). Interventions to improve people’s deepfake detection yielded mixed results. Humans and AI-based detection models focus on different aspects when detecting, suggesting a potential for human–AI collaboration. The findings highlight the complex interplay of factors influencing human deepfake detection and the need for further research to develop effective strategies for deepfake detection.

When a person chooses to interact with their phone instead of their partner (e.g., technoference, phubbing), it may diminish interactional quality, relationship satisfaction, and well-being. However, much of the research on technology use in relationships has utilized self-reports. We extend prior work by objectively measuring smartphone use in a sample of 247 adult participants (75% women; mean age = 30.87 years) to better understand the extent of use around one’s partner and the connection between this use and relational and personal well-being. Participants completed an online baseline survey and 8 days of phone tracking and nightly time diaries. On average, participants used their smartphone during 27% of their time around their partner; 86% used their phone every day at least some around their partner. Linear regression modeling revealed that phone use around partner (not total daily phone use) predicted lower relationship satisfaction and coparenting quality, although effects were only significant for women. We also found that phone habits in general (i.e., both phone use around partner and total phone use) predicted greater depression and lower life satisfaction, with effects trending toward being stronger for women. Overall, our results suggest that one’s own phone use is connected—especially for women—to one’s own relational and personal well-being. Our objective phone use and daily diary methods offer one potential model for studying the nuances of technoference and its effects on relational and personal well-being. Future research should continue to explore both objective and subjective measures of device use within couples and families.

Telepractice in neuropsychology has become increasingly prevalent in recent years due to its ability to provide accessible and convenient care to patients regardless of their location. However, the validation of many neuropsychological tools for distance assessments remains limited, and there is a particular lack of remotely administered assessment tests with alternate forms, which are crucial for monitoring symptoms and performance in clinical contexts and for minimizing practice effects in research practice. Consequently, the present study was aimed at evaluating the consistency of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) scores across videoconference and face-to-face administrations and to examine whether the scores obtained via videoconference support interpretations similar to those obtained via face-to-face administration. A total of 185 participants aged between 20 and 79 years (M = 46.24, SD = 19.63) underwent RBANS testing twice: once in person using the standard pen-and-paper modality and once remotely via videoconference, using Alternate Forms A and B to mitigate the learning effects. Results from the linear mixed models revealed no significant differences between remote and face-to-face administrations based on the modality of administration (p > 0.05). Bayes factors supported the null hypothesis, suggesting that RBANS performance is consistent across the two modalities of administration. However, discrepancies were observed in certain subtests between alternate forms of the RBANS, highlighting the need for standardization. In conclusion, findings suggested that the same norms that are used to interpret the RBANS scores obtained via face-to-face administration may be employed when administered remotely through videoconferencing. Accordingly, the study provides valuable insights into the feasibility of remote neuropsychological assessment and underscores the potential utility of videoconference technology in clinical and research settings.

Emotion recognition in conversations (ERC) is of high importance, especially when it relates with human behavior assessment. Nevertheless, ERC so far has mainly focused on the identification of each interlocutor’s emotions. Here, for the first time, we consider the concept of emotion climate (EC), that is, the emotion reciprocally established by the peers during a naturalistic conversation, and we introduce machine learning (ML) models that efficiently perform emotion climate recognition (ECR). The latter is explored in the cases where the EC is (a) perceived within a conversational group, (b) conveyed from interlocutors involved in a conversation to the external observers, and (c) felt by the external observer. Features from conversational speech and affect dynamics (AD) data (n = 4685), drawn from three open datasets (i.e., K-EmoCon, IEMOCAP, and SEWA), were inputted to the ML-based ECR, achieving maximum accuracy of 96% and 83% in the K-EmoCon and IEMOCAP datasets, respectively. Cross-lingual validation was performed on SEWA dataset, justifying the generalization potential of the proposed approach. These results show that efficient ML-based ECR can identify how the EC is jointly built, perceived, and felt by others, providing a new approach in assessing emotional aspects in naturalistic conversations.

The immersion of artificial intelligence (AI) in higher education presents significant challenges and opportunities. This study examines the acceptance of AI as a teaching strategy among university teachers, following the extended UTAUT2 model with the inclusion of the teacher skills and knowledge for technology integration (SKTI) construct. Employing a quantitative cross-sectional research design, data were collected from 318 university teachers with prior experience using AI as a learning strategy through nonprobabilistic convenience sampling across 10 universities in northern Peru. Participants completed an online survey, and data were analyzed using descriptive statistics, Kruskal–Wallis tests with Dunn’s post hoc comparisons, and partial least squares structural equation modeling (PLS-SEM). The results showed that performance expectancy (β = 0.129^∗∗), hedonic motivation (β = 0.167^∗∗), habit (β = 0.405^∗∗∗), and SKTI (β = 0.263^∗∗∗) had a positive influence on the behavioral intention to adopt AI as a teaching strategy. Additionally, behavioral intention (β = 0.303^∗∗∗), facilitating conditions (β = 0.115^∗), and habit (β = 0.464^∗∗) determine the behavioral use of AI by teachers. The Kruskal–Wallis test revealed significant differences among age groups in the performance expectancy, social influence, habit, and behavioral intention constructs, with the 37- to 48-year-old age group showing higher average ranks. The discussion highlights that these findings suggest a positive adoption of AI among teachers, driven by individual and contextual factors, and challenges assumptions about the relevance of certain constructs in this specific context. In conclusion, this study represents a significant advancement in understanding the adoption of AI in university teaching and provides valuable guidance for practical implementation efforts.

Cervical cancer presents a significant global health challenge, affecting patients and healthcare systems. Early identification and accurate prediction of risk factors are essential for reducing incidence and improving patient outcomes. This study focuses on predicting indicators and diagnosing cervical cancer using a comprehensive dataset that includes demographic information, lifestyle factors, and medical histories. We developed a predictive model to aid early diagnosis and identify key risk factors. The dataset consists of four cervical cancer tests—Hinselmann, Schiller, cytology, and biopsy—with 858 participants and 30 features. We addressed 22.14% of missing values using the MICE iterative imputer and balanced the data through the synthetic minority oversampling technique (SMOTE). We applied five machine learning algorithms: random forest (RF), linear regression (LR), support vector machine (SVM), K-nearest neighbors (KNN), and extreme gradient boosting (XGBoost). The SpFSR technique was utilized to enhance feature selection, assessing how a subset of features could maintain high accuracy compared to the full model. Our findings showed that selecting fewer features, such as half or even a quarter of the variables, still yielded strong results, emphasizing the importance of careful feature selection in cervical cancer prediction. The RF algorithm achieved the highest accuracy, with 99% using the full feature set and 98% with a reduced set of five features. Notably, diagnosis and hormonal contraceptives were identified as significant predictors. Hormonal contraceptives, which can affect cervical health, are linked to increased risks of HPV infection and cervical cancer. This study highlights the role of SpFSR in improving prediction models and suggests that external validation is necessary to confirm our findings in diverse populations. Further research should explore additional datasets and variables not covered in this study, as well as the model’s practical applicability in clinical settings.

Behavioral science confronts the issue of how people’s behaviors differ from what they intend to do. However, current models, such as the theory of planned behavior, are insufficient to account for contextual influences and interdisciplinary effects, especially in the case of modern social phenomena. The majority of studies concentrate on single domains (e.g., health and consumer behavior) and employ manual coding schemes, overlooking essential thematic relationships. This research highlights the necessity for integrative frameworks that attempt to analyze why intentions fail to be realized in complex settings such as climate change and digitalization. The primary objectives of this research are to identify and operate dominant and emerging thematic trends in intention–behavior literature in a time series from 1979 to 2025 and to analyze and investigate the effects of publication index status and citation patterns on scholarly impact. This study uses structural topic modeling (STM) alongside bibliometric analyses to identify themes and correlations in intention–behavior research. STM employs generalized linear models to include document-level metadata, allowing for the discovery of related topics and the key factors influencing the development of the literature. Data collection was initially performed on February 20, 2025, through the Web of Science database, using studies that were identified following PRISMA guidelines, reviewed, and considered relevant. The initial records numbered 5350. Significant thematic trends were found to define, and key psychological mechanisms to explain the intention–behavior gap were identified. The study also found that the determinants of publication index status and citation trends play important roles in establishing the discipline’s fate and the impact of intention–behavior literature. Based on these findings, the study highlights how strong thematic links in intention–behavior research can inform cross-domain interventions—such as integrating physical activity and organic food campaigns or leveraging sustainable tourism to promote ethical consumption—by targeting shared psychological drivers like health identity and self-image. In future research, the intention–behavior gap should be investigated across different disciplines and contexts and with longitudinal and experimental designs to take advantage of the psychological and contextual factors that affect behavior.

The acceptance and adoption of emerging technologies are crucial for their effective integration. This study examines the factors influencing educators’ acceptance of Generative AI (Gen AI) tools in higher education, guided by the UTAUT model. It also develops a structural model to explore the relationships between UTAUT constructs and behavioral intention (BI) to use Gen AI. Using a quantitative approach, the study collected data through a self-administered online survey based on prior research findings. The survey gathered responses from 307 educators across various Arab countries who are early adopters of Gen AI in teaching. PLS-SEM was used to analyze the data. Findings indicate that UTAUT constructs significantly and positively influence educators’ intention to use Gen AI. Additionally, the results highlight the complex role of gender and work experience, revealing diverse perspectives among educators from different countries. This study contributes to the literature by deepening the understanding of technology adoption factors. It also offers theoretical and practical implications for researchers and policymakers in designing strategies to integrate Gen AI into higher education in developing countries.

Climate change is a global phenomenon that affects every living being on our planet. Raising awareness among people about climate change and helping them realize the possible consequences of their actions is key to mitigating climate change problems. Our research was aimed at achieving this by building a persuasive intervention that combines visualization of climate change data and an interactive narrative that demonstrates how our actions can impact the climate. We conducted a user study with 100 participants and found evidence showing that our system was effective in significantly promoting behavioral intention to mitigate climate change. We found defensive responses as a key factor that is negatively influencing the effect of our intervention on the participants. Compelling visuals and multiple interaction options, simulating climate actions and their consequences, and reducing the effort to learn about the phenomenon were significant positive techniques used in the intervention. Additionally, the social elements of our intervention played a major role in promoting participants’ willingness to perform proenvironmental behavior. Our work contributes to the field of persuasive technology, data visualization, interactive narratives, and climate research by introducing a new persuasive way of communicating climate change information to the public using a combination of data visualizations and interactive narratives.

The adoption of blockchain technology continues to grow, a direct result of its potential to provide new solutions to old problems in several industries, including the electoral sector. Blockchain technology is proposed to have the potential to address and overcome the traditional pen and paper scheme’s challenges and limitations, as well as trust concerns around more modern e-voting systems. Ultimately, with the aim to revert the recent downward trend in voter turnover, despite the interest and potential, there remains a significant research gap in understanding citizen response to this technology. This research is aimed at investigating whether citizens would be willing to embrace blockchain technology, as well as at exploring the factors that influence its adoption. A model designed to combine the extended unified theory of acceptance and use of technology methodology with an experimental approach is applied. The results of the study (N = 416) show that the intention to use blockchain-based e-voting systems can be predicted by five of seven constructs, that is, citizens are more likely to adopt e-voting systems when they perceive them to be effective, socially endorsed, enjoyable, trustworthy, and low in perceived risk. However, we do not find a direct influence of blockchain technology, over cloud-based e-voting, on voting intentions indicating that the benefits of this approach may not be well understood by consumers or may not drive the desired increase in voting intention. By understanding citizens’ willingness and concerns to adopt new voting technologies and the factors influencing this disposition, policymakers are better equipped to develop strategies on the development and implementation of electronic voting systems and can make informed choices about the use of blockchain technology.