The rapid spread of misinformation on digital platforms has made it increasingly difficult for users to identify reliable news. Most existing news platforms focus only on aggregating content and do not provide built-in mechanisms to verify the credibility of the information they present. This creates a gap where users consume news without knowing whether it is trustworthy or not. To address this issue, this paper presents News Garden, an AI-powered interactive 3D globe news explorer that combines news aggregation, visualization, and fake news detection into a single platform. The system gathers real-time news from multiple sources such as GNews API and GDELT, ensuring broad and diverse coverage across different regions. Each article is analyzed through a multi-dimensional process that includes sentiment detection, category classification, credibility scoring, entity extraction, geographic mapping, and automatic summarization. For credibility assessment, the system uses an ensemble approach that combines a RoBERTa-based model with Google Gemini 2.0 Flash, allowing it to evaluate news more accurately by considering both textual patterns and contextual understanding. An interactive 3D globe interface enables users to explore news geographically, making the experience more intuitive and engaging. Additional features such as a News Verifier tool, analytics dashboard, and dedicated India state-level news coverage further enhance usability and provide deeper insights. The system also highlights trending news using a ranking mechanism based on recency, credibility, and sentiment. Overall, the proposed system offers a practical and user-friendly solution to reduce the impact of misinformation by integrating AI-based credibility analysis directly into the news browsing experience, helping users make more informed decisions.

TikTok has become an important information source for young adults, yet little is known about how they assess the credibility of content encountered on the platform. This study builds on a combination of a 7-day diary study with semistructured qualitative interviews with 46 young adults (aged 18–24), providing nuanced insights into credibility assessments during everyday platform use. The findings reveal that participants are generally skeptical of TikTok’s overall credibility, but rely heavily on source authority as a heuristic for judging individual posts. Message characteristics play largely supportive roles, as TikTok’s affordances—particularly its entertainment orientation and aesthetics—make many cues ambiguous. Participants predominantly employ internal verification strategies, drawing on their own knowledge and intuition, and use external strategies only occasionally, mainly by checking on-platform comment sections. Our findings challenge the applicability of traditional credibility models and demonstrate the importance of considering platform affordances when studying credibility assessments.

Non-suicidal self-injury (NSSI) persists as a major public health challenge worldwide. Identifying and strategically targeting risk factors for NSSI constitutes a practical approach to its prevention. We aim to synthesize existing knowledge concerning the range and magnitude of risk factors for NSSI among children and adolescents, and to critically assess the robustness of the available evidence. In this umbrella review, six bibliographic databases were systematically searched for articles published from database inception to Dec 2024. For the assessment of evidence credibility, pre-specified criteria for classifying evidence were utilized, categorized as convincing ("class I"), highly suggestive ("class II"), suggestive ("class III"), weak ("class IV"), or no evidence ("class V"). The Amstar-2 framework was employed to evaluate the quality of the evidence which graded as "high," "moderate," "low," or "critically low" quality. The study included meta-analyses of observational studies in the past 30 years on risk factors for NSSI in children and adolescents. We identified 16 meta-analyses comprising 410 primary studies on 43 risk factors from 38 countries, involving 2,659,156 children and adolescents. Twenty-three (e.g. LGBTIQ) risk factors were categorized as individual, followed by family level (n = 8, e.g. childhood maltreatment), school/peer level (n = 8, e.g. bully victims) and multifactorial level (n = 4, e.g. no religion). Eighteen (41.86%) risk factors provided highly suggestive (Class II) evidence of association with NSSI. Suggestive evidence (class III) indicated that NSSI was associated with adverse childhood experiences (2.31, 1.77-3.01) and being left-behind children (1.37, 1.11-1.69). A multitude of risk factors spanning diverse domains were identified, highlighting the multifactorial nature of NSSI in adolescents and children. Comprehensive prevention strategies and measures should be conducted for children and adolescents to decrease the risk of NSSI and associated harms in multilevel approaches.

Neural networks have been extensively applied in mechanical fault diagnosis due to their strong capabilities in feature extraction and classification. However, their limited interpretability and unknown credibility of decision hinder deployment in high-reliability scenarios. To address this issue, a frequency band multi-indicator feature embedding (FIE) method based on physical information embedding is proposed. A filter bank integrated into the convolutional layer is employed to simulate the spectrum, upon which multifrequency domain indicators are computed to extract multichannel physical features that vary continuously with frequency bands. A class-aware weight mask, enabling interclass distribution differentiation, is generated to assign distinct channel weights for different faults. This facilitates the extraction of key features for decision analysis and enables credibility evaluation based on a distance metric. Experimental results on multiple fault datasets demonstrate that the FIE module exhibits strong noise robustness and enhances diagnostic performance under variable-speed conditions. Furthermore, the class-aware mask supports reliable credibility assessment and feature interpretation, thereby supporting feature-level interpretation and decision credibility assessment while maintaining internal transparency.

Abstract Nigerian citizens fleeing insurgent violence or anti- LGBTQ + laws form a growing cohort of U.S. asylum applicants, yet their claims often turn on credibility assessments shaped by evidentiary gaps. Digital evidence, social media activity, messaging records, and online traces, offer new avenues for corroboration but also introduce risks of misinterpretation, privacy intrusion, and structural bias, particularly for individuals who conceal their identities to survive. This article investigates the U.S. legal framework governing credibility determinations under the REAL ID Act and key decisions such as Matter of A-B- , analysing how digital expectations intersect with the lived realities of applicants fleeing Boko Haram and persecution under Nigeria’s Same Sex Marriage (Prohibition) Act of 2013. Drawing on international human rights standards, including the UNCHR Handbook and the ICCPR , it argues that digital evidence has become a decisive, and sometimes distorting, force in asylum adjudication and offers reforms to ensure more context-sensitive and rights-consistent credibility assessments.

Neurodivergent people, including individuals with intellectual and developmental disabilities, continue to face systemic barriers to meaningful and fair participation in the U.S. justice system. Legal standards governing competence, credibility, and culpability remain anchored in expectations of neurotypical communication and reasoning. These expectations do more than shape procedures; they define who is heard, believed, and ultimately brought to justice. This commentary examines forensic ableism, the privileging of neurotypical cognition and communication in legal contexts, through Fricker's framework of epistemic injustice, with a focus on testimonial injustice. In practice, credibility judgments are rooted in neurotypical norms that often devalue neurodivergent testimony. Across competency evaluations, credibility assessments, and capital sentencing decisions, disability-linked patterns of expression and interaction are frequently misinterpreted as signs of unreliability or diminished competence. Addressing forensic ableism requires the redesign of legal processes and broadened disability education to aid in the recognition of diverse cognitive and communication profiles as legitimate ways of knowing and participating. We call for reforms grounded in accessibility, epistemic humility, and collaboration with the neurodivergent community.

Does the source and distinctiveness of misinformation in social media posts affect its believability? Across two studies, we examine this question utilizing vignette experimental survey designs that manipulated the source, accuracy of information, and whether the information was familiar or distinctive (unfamiliar). Four different topics with different political or moral positions were used to assess how effects varied across topics. True posts were found to be more believable than misinformation posts across the four topics in Study one (n = 595). In study two (n = 514), misinformation was rated as more believable, more accurate, and more trustworthy if it was unfamiliar rather than familiar. Source effects were significant but smaller than the distinctiveness effect. Posts from the source of authority were rated as more believable than those from a friend. Distinctive messages receive initial assessments of higher credibility, suggesting a heuristic process. However, the personally relevant topic of COVID, showed higher believability for unfamiliar misinformation, but also a higher percentage intending to verify the information through additional research. Those supporting more conservative views perceived misinformation as more believable. These findings are consistent with the Elaboration Likelihood Model of Persuasion. Implications and directions for future research are discussed.

The EQ‑5D is widely applied to measure patient-reported outcomes, yet its minimally important difference (MID) has not been clearly established across distinct recovery phases after stroke. This study aimed to determine phase‑specific MIDs in EQ‑5D following stroke and to explore heterogeneity by estimation method, direction of change, and stroke etiology. A total of 9978 adults with neuroimaging‑confirmed acute ischemic stroke were included in a prospective longitudinal cohort study. EQ‑5D and modified Rankin Scale (mRS) scores were recorded at admission (V1), hospital discharge (V2), 3-month (V3), and 1-year since admission (V4). Anchor-based MIDs were estimated at both group and individual levels and triangulated by distribution-based and instrument-defined approaches. Changes during the recovery phases (V1–V2, V2–V3, and V3–V4) were grouped into 3 categories: improved, no change, and deteriorated. Subgroup analyses were conducted according to the TOAST classification. Credibility of MID estimates was assessed using a validated instrument for anchor-based methods. Phase-specific group-level MIDs for improvement decreased over time: anchor-based estimates were 0.19 at V2, 0.14 at V3, and 0.11 at V4, while deterioration MIDs were smaller. Distribution-based and instrument-defined estimates fluctuated slightly around the anchor-based values but followed a similar downward trend over time. Individual‑level analyses yielded MIDs with acceptable discriminative power (area under the curve ≥ 0.70) only for improvement at V2 (0.10) and V3 (0.01). Cardioembolic strokes had higher MIDs than large-artery atherosclerosis and small-artery occlusion, while baseline utilities showed the reverse. Credibility assessment confirmed high reliability. This study provides the phase‑specific MIDs for utility measures after ischemic stroke, showing a declining trend from acute to chronic recovery and confirming robustness across multiple estimation methods. While group‑level MID ranges are recommended for effect size interpretation, trial design, and evidence certainty rating, individual‑level thresholds should be reserved for personalized evaluation. These values assist in the interpretation of patient‑reported outcome changes and evaluation of healthcare interventions across different recovery phases. ClinicalTrials.gov NCT02470624. Registered 10 June 2015.

Eye movements reflect neurocognitive and psychophysiological processes linked to attention, cognitive load, and emotional arousal, making them relevant for forensic investigations. This study examined the integration of eye-tracking with Virtual Reality (VR) crime scene exposure. Forty participants (20 experimental, 20 control) were exposed to a VR mock crime scene and assessed using the Eye Detect Polygraph. The experimental group showed significantly greater fixation duration, gaze activity, pupil dilation, and AOI engagement for crime-related stimuli (p<.05). Eye Detect results largely classified experimental participants as deceptive and controls as credible or inconclusive, supporting VR-based eye tracking and ocular credibility assessment as a potential preliminary screening tool in forensic psychological investigations.

In courtroom interaction, the pervasive assumption that “seeing is believing” grants visual conduct, especially gazing behavior, a crucial role in shaping perceptions of testimonial credibility. While prior legal and psychological research has acknowledged that gaze may influence credibility judgments, little is known about how litigants themselves actively mobilize gaze as a rhetorical resource during direct evidential presentation, particularly within the interactional complexity of authentic trials. Drawing on multimodal conversation analysis, this study conducts a fine-grained examination of gazing behavior in coordination with co-occurring verbal and embodied resources. Based on video data from the 2022 Depp v. Heard defamation trial, the analysis identifies three functional gazing categories and integrates the Aristotelian rhetorical triad to illustrate how litigants strategically mobilize visual conduct to fulfill distinct persuasive objectives: evaluative gazing grounds logos to reinforce logical coherence; regulatory gazing projects ethos to display institutional legitimacy; affective gazing embodies pathos to solicit emotional empathy. These findings demonstrate that testimonial credibility is an interactional accomplishment constructed through the interplay of verbal accounts and embodied conduct. By foregrounding gaze as a multimodal rhetorical resource, this study contributes to multimodal discourse and legal rhetoric research by offering an empirically grounded framework for analyzing how credibility is visually enacted in courtroom testimony, with implications for both scholarly analysis and judicial credibility assessment practices.

ABSTRACT In sexual abuse cases, the witness statement plays a fundamental role. This brings about the complex task for judges to evaluate the statement on its credibility. In Study 1, we polled 79 Dutch legal professionals about the criteria they rely on to evaluate the credibility of a statement. Most criteria mentioned pertained to the content of the statement, specifically consistency (66%), accuracy (66%), and detailedness (53%). Twenty-eight percent mentioned all three criteria. The way the statement is presented was also mentioned (non-verbal behaviour: 14%, emotion; 13%; authentic impression: 11%). In Study 2, we analyzed 518 Dutch court rulings on sexual assault on the criteria used in credibility assessment. The most often used criteria were again consistency (80.1%), detailedness (65%), and accuracy (31%), with 13% relying on all three criteria. Authentic impression (23%), emotionality during the statement (13%), and signs of trauma (9.5%) were also regularly used. In conclusion, legal professionals rely heavily on content criteria (Consistency, Accuracy, and Detailedness) – but not in a systematic way. Highly contested criteria (i.e., emotion, trauma, nonverbal behaviour, authentic impression) are also regularly used. Judicial decision-making may benefit from relying exclusively on validated indicators and doing so in a more systematic manner.

Surgical video-based learning has rapidly shifted to unregulated open-access platforms. This study evaluates the consumption habits, clinical impact, and content selection criteria of urologists across different career stages to understand the risks of this transition. A cross-sectional survey was conducted among 133 urologists stratified as Residents, Specialists, and academic urologists. Platform preferences, practice modifications, and selection criteria were assessed using Likert scales. Differences between groups were analyzed using Chi-square and Kruskal-Wallis tests with post-hoc pairwise comparisons. Open-access platforms (YouTube) were the primary source for 84.0% of the cohort, with no significant difference between groups (p=0.351). A total of 37.4% reported modifying their surgical technique based on online videos. This rate did not differ significantly between residents (34.9%), specialists (48.0%), and academic urologists (35.0%) (p=0.477). In adjusted analysis, career stage was not independently associated with technique modification, whereas higher weekly operative workload was associated with a greater likelihood of reporting modification. However, a significant divergence emerged in selection criteria: Residents prioritized "Surgeon's Reputation" significantly less than Specialists (p=0.012) and Academic urologists (p=0.007). Additionally, the clarity of patient consent in online videos was rated universally low. The "YouTube era" appears to have a tangible influence on clinical practice across different seniority levels. However, potential differences in credibility assessment may exist, as residents may rely more heavily on unregulated content without systematically evaluating author expertise. These findings highlight the potential value of incorporating digital literacy and source evaluation skills into residency curricula to support more informed engagement with online surgical content.

Despite the fact that modeling and simulation are now recognized as promising innovative methodologies, their use in the context of development of drugs for sickle cell disease and Thalassemia has not yet been reviewed. Considering the challenges of conducting clinical trials for hemoglobinopathies, our work aims at exploring the current status of use of modeling and simulation by drug developers and their endorsement by regulators, based on European Medicines Agency scientific advices given to industry by the Committee for Medicinal Products for Human Use from January 2000 to December 2024. The present study describes different aspects of how modeling and simulation methods have been used and assessed. Our results highlight the need for an openly accessible structured framework that details the standards/criteria that make a method acceptable for regulators. The use of the ICH M15 credibility assessment framework is suggested for this purpose. Bearing in mind the question-centric approach, a repository of drug development questions and associated methods and data was built from 45 and 31 scientific advices for sickle cell disease and Thalassemia, respectively. The classification of the questions, methods and data based on the question-centric approach enabled modeling and simulation methods to be easily identified, objectively assessed and benchmarked against alternative methods used to address the same questions to evaluate their performance for characterizing benefit/risk of the different drugs.

The International Journal of Interactive Multimedia and Artificial Intelligence (IJIMAI) provides a forum for researchers and professionals to share recent advances in artificial intelligence (AI) and its wide range of applications. This issue brings together contributions that reflect the growing diversity of AI research, covering topics such as information credibility assessment in social media, recommender systems, federated learning and data privacy, medical image analysis, educational technologies, large language models and virtual assistants, intelligent multimedia systems, and environmental monitoring. Collectively, these works illustrate how current AI methods continue to expand across disciplines, addressing both theoretical challenges and real-world problems.

This opinion paper examines the transformative relationship between Generation Z and business information, arguing that this cohort is pioneering a bottom-up, information-driven approach to economic self-reliance. Unlike previous generations, Gen-Z leverages a diverse digital ecosystem from social media platforms to specialized forums to gather, synthesize, and apply business intelligence for career advancement and venture creation. While this represents a significant democratization of entrepreneurial opportunity, it also presents critical challenges, including information overload, credibility assessment, and a potential foundational skills gap. This paper discusses these promises and perils, contending that the current institutional support systems are lagging behind these informal learning models. The conclusion offers a multi-stakeholder call to action, proposing that educators, policymakers, and business leaders must collaboratively build a scaffolded ecosystem to effectively harness Gen-Z’s unique information sensibilities for sustainable economic growth.

ABSTRACT This article explores the politics of truth and recognition in the assessment of queer asylum claims in Denmark, focusing on how credibility is produced, evaluated and governed through affect. Drawing on ethnographic fieldwork and interviews with queer asylum seekers and immigration officers, alongside rejection letters from the Danish Immigration Service, I explore how asylum seekers are compelled to produce intimate affective archives of desire, fear and identity. These archives function both as forms of evidence and as sites of state discipline, yet they often exceed bureaucratic logics, appearing in ambivalent, fleeting and resistant forms. This article shows how the circulation and evaluation of affect operates as a technology of governance that shapes inclusion, exclusion and deservingness within the asylum system. By foregrounding ambivalence in queer affective archives, it argues for alternative epistemologies of queerness that challenge normative assumptions of coherence and consistency in credibility assessments.

The efficacy of asciminib was proven in newly diagnosed (first-line, 1L) patients with Philadelphia-positive chronic myeloid leukemia in chronic phase (Ph+ CML-CP) and in patients treated with at least two prior tyrosine kinase inhibitors (third-line, 3L+). Given that no randomized controlled trial has been conducted for second-line (2L) patients with CML, this analysis aims to infer the efficacy of asciminib in the 2L setting using the 80 mg once-daily dosing regimen and to support the use of asciminib in patients with CML-CP irrespective of line of therapy. Data (n = 430) were used from three studies, including first-in-human and ASCEMBL in 3L+ and ASC4FIRST in 1L trials, to evaluate the effect of line of therapy on efficacy on the basis of the time-course of BCR::ABL1 mRNA transcripts. Previously adapted to characterize the effect of nilotinib as a 1L and 2L therapy on a CML surrogate marker, the model has three compartments representing quiescent leukemic stem cells and proliferating drug-susceptible and resistant bone marrow cells, wherein the effect of asciminib was modeled as an enhancement of the elimination of susceptible cells. Asciminib efficacy in 2L was inferred from 1L by borrowing information from nilotinib data. A credibility assessment showed robust prediction accuracy and precision of the model with external 2L data from the ASC2ESCALATE study (n = 36). Major molecular response (MMR) rates in 2L were predicted to be 61-67% at week 48 and 70-76% at week 96, with the 80 mg total daily dose. A model-informed drug development (MIDD) approach was applied to predict 2L efficacy and supported global regulatory approval of asciminib across treatment lines, despite limited clinical data in 2L.

Purpose With the rapid development of AI technologies, the problem of multimodal disinformation (MD) has become increasingly serious on social media, posing significant threats to various facets of human society. Although algorithm-based detection could be a solution, human-oriented verification is still crucial in combating online multimodal disinformation. However, there are empirical and theoretical gaps that require attention. Hence, this study investigates human-oriented verification cues and underlying information processing patterns. Design/methodology/approach Guided by the Heuristic-Systematic Model and established information credibility assessment frameworks, we conducted in-depth interviews with 24 social media users for data collection. Subsequently, a hybrid deductive and inductive thematic analysis approach was adopted for qualitative data analysis. Findings We uncovered five key categories of verification cues that were leveraged by individuals, namely (1) modality, (2) content, (3) source, (4) engagement and (5) external. Within these categories, a total of 11 cues were further identified. Additionally, we revealed three information processing patterns underlying the employment of these cues, namely (1) systematic, (2) heuristic and (3) dual-process verification. Generally, our findings show that in social media environments, individuals strategically utilize multiple informational and social cues for MD verification. In this process, dual-process verification is necessary for consolidating judgment-making, enabling the mutual complement between systematic and heuristic cues. Originality/value Distinct from existing research on algorithm-based mitigations, this study adopts a human-oriented perspective. Our proposed categorization of verification cues and patterns serves as a framework for evaluating and validating MD in social media, thereby laying a foundation for further research. Practical implications are also discussed.

Purpose Driven by the exponential growth in the online engagement of consumers, it has become essential for brands to promote their products and services through online influencer marketing (OIM) campaigns. OIM stems from the integrated marketing communications (IMC) theory, and in addition to the benefits, OIM campaigns pose certain challenges for brands. Therefore, this study identifies and prioritizes the OIM challenges from the lens of the IMC theory. Design/methodology/approach The author examined the literature and identified 12 challenges in the context of the IMC theory. Following the collection of feedback from 22 industry professionals, the highlighted challenges were ranked using a unique causal impact analysis technique called the grey influence analysis (GINA). Findings The study revealed that negative brand association risk, influencer authenticity and credibility assessment, cost and return-on-investment management, and campaign performance management are the most influential individual OIM challenges. Whereas a collective categorical analysis revealed that the challenges pertaining to audience-centered and channel-centered IMC pillars are the most significant. Research limitations/implications The study focused on OIM campaigns in general, rather than any specific industry. Consequently, the findings may not be replicable in specialized industries. Practical implications The study’s findings could help marketing managers boost OIM campaign execution in line with the IMC by focusing on brand alignment, credibility evaluation, and performance accountability of online influencers. The study encourages ethical and transparent influencer practices through a responsible online communication system. Originality/value As one of the first applications of GINA in OIM, this study broadens its scope to the wider domains of digital marketing and branding.

Credibility assessment is a central component of refugee status determination (RSD), yet its relationship to the enactment of injustice—particularly with respect to knowledge production—remains insufficiently examined. This article investigates how credibility assessments marginalise asylum seekers as subjects of knowledge and contribute to epistemic injustice. Drawing on epistemic injustice theory, the analysis is based on semi-structured interviews with decision-making actors in Brazil. The findings suggest that refugee status determination is not merely declaratory, but is constituted through epistemic practices that systematically marginalise applicants’ knowledge, making the granting or denial of asylum contingent upon the enactment of epistemic injustice.