Popular Apps Research Articles

Demystifying React Native Android Apps for Static Analysis

React Native, an open-source framework, simplifies cross-platform app development by allowing JavaScript-side code to interact with native-side code. Previous studies disregarded React Native, resulting in insufficient static analysis of React Native app code. This study initiates the investigation of challenges when statically analyzing React Native apps. We propose ReuNify to improve Soot-based static analysis coverage for JavaScript-side and native-side code. ReuNify converts Hermes bytecode to Soot’s intermediate representation. Hermes bytecode, compiled from JavaScript code and integrated into React Native apps, possesses a unique syntax that eludes current JavaScript analyzers. Additionally, we investigate opcode distribution and conduct in-depth analyses of the usage of opcode between popular apps and malware. We also propose a benchmark consisting of 97 control-flow-related cases to validate the control-flow recovery of the generated intermediate representation. Furthermore, we model the cross-language communication mechanisms of React Native to expand the static analysis coverage for native-side code. Our evaluation demonstrates that ReuNify enables an average increase of 84% in reached nodes within the call graph and further identifies an average of two additional privacy leaks in taint analysis. In summary, this paper demonstrates that ReuNify significantly improves the static analysis for the React Native Android apps

Ethical implications of artificial intelligence in skin cancer diagnostics: use-case analyses.

Skin cancer is the most common cancer worldwide. Early diagnosis is crucial for improving patient survival and morbidity. Artificial intelligence (AI)-assisted smartphone applications (apps) for skin cancer potentially offer accessible, early risk assessment of suspicious skin lesions. However, the integration of novel technologies into dermatology pathways raises ethical concerns. Although ethical principles for AI governance are well known, how these principles should be applied to real-life AI apps readily available for public use is less well understood. We conducted an ethical use-case analysis of commercially available skin cancer apps to better understand the ethical issues arising from their development and use in a real-world context. Established methods for ethical analysis of clinical AI applications were applied to two popular skin cancer apps in the UK: SkinVision and Scanoma. Systematic searches of published literature, regulatory documents, and websites were conducted to review the evidence regarding app development, effectiveness, and use. Screening for inclusion was undertaken by two researchers independently. Ethical concerns were identified with reference to previously described ethical concerns and principles for AI-assisted healthcare. By conceptualising ethical principles within the use-context of skin cancer apps, we identified specific ethical issues arising throughout the AI lifecycle of both apps. One company provided extensive detail regarding algorithm development and decision-making, this information was insufficiently reported for the other app. Other concerns identified related to number, quality, and consistency of studies assessing algorithm efficacy. Limited efforts to address potential skin tone biases and exclusion of individuals with darker skin tones as target users by one app risks perpetuating existing inequalities. Inadequate regulatory oversight was identified. Findings from our ethical use-case analysis of two patient-facing AI-assisted skin cancer apps suggest inadequate incorporation of bioethical norms such as justice, responsibility and transparency into the development and deployment of both apps. Improved regulation should increase accountability. Ensuring ethics by design through integration between technology developers, dermatologists, ethicists, and the public is urgently needed to prevent the potential benefits of AI-assisted skin cancer apps being overshadowed by potential ethical harms.

Filtering Useful App Reviews Using Naïve Bayes—Which Naïve Bayes?

App reviews provide crucial feedback for software maintenance and evolution, but manually extracting useful reviews from vast volumes is time-consuming and challenging. This study investigates the effectiveness of six Naïve Bayes variants for automatically filtering useful app reviews. We evaluated these variants on datasets from five popular apps, comparing their performance in terms of accuracy, precision, recall, F-measure, and processing time. Our results show that Expectation Maximization-Multinomial Naïve Bayes with Laplace smoothing performed best overall, achieving up to 89.2% accuracy and 0.89 F-measure. Complement Naïve Bayes with Laplace smoothing demonstrated particular effectiveness for imbalanced datasets. Generally, incorporating Laplace smoothing and Expectation Maximization improved performance, albeit with increased processing time. This study also examined the impact of data imbalance on classification performance. Our findings suggest that these advanced Naïve Bayes variants hold promise for filtering useful app reviews, especially when dealing with limited labeled data or imbalanced datasets. This research contributes to the body of evidence around app review mining and provides insights for enhancing software maintenance and evolution processes.

Vaccine Hesitancy and COVID-19 Risk Behaviors Associated with Social Media Use in Japan

ABSTRACT We examined the associations between the use of different types of media and COVID-19 vaccine hesitancy, as well as risk behaviors of COVID-19 infection, in Japan in late 2021. Cross-sectional surveys were conducted using rapid online surveys of residents in Iwate Prefecture from February 5 to 7, 2021, and from October 1 to 3, 2021. Each individual’s risk of acquiring SARS-CoV-2 infection was calculated using a quantitative assessment tool (the microCOVID). Intention to get vaccinated for COVID-19 was assessed by self-report. Usage of five types of media for obtaining COVID-related information was assessed: (1) newspapers, (2) television or radio, (3) internet or news apps, (4) social network services (SNS) (excluding LINE, a popular messaging app), and (5) other. Reliance on SNS did not show significant associations with either intention to get vaccinated or engaging in risky behavior for acquiring COVID-19. Although users of the internet or news apps were marginally significantly less likely to engage in high-risk behavior, significant associations between vaccine hesitancy and the usage of the internet or news apps were found in the middle age and elderly groups (OR [95% confidence interval (CI)] in middle age: 1.55 [1.07–2.23]; in elderly; 9.24 [3.28–26.02]). The differential associations between different types of media use and COVID-19 prevention behaviors may assist in preparing for future pandemic outbreaks. One implication for public health risk communication is audience segmentation, such as emphasizing vaccine safety and effectiveness for older audiences.

Improving the Science of Adolescent Social Media and Mental Health: Challenges and Opportunities of Smartphone-Based Mobile Sensing and Digital Phenotyping

AbstractThe impact of social media (SM) use (‘screentime’) on adolescent mental health has been the focus of increasing concern, despite mixed findings from empirical research. Current methodological approaches rely on self-reported SM use, which has limited accuracy and obscure the dynamic interplay of SM use and mental health. Smartphone-based mobile sensing offers new opportunities to gain insights into adolescents’ SM use patterns and behaviors, particularly at an idiographic level. Considerations and challenges of smartphone sensing methods for capturing adolescents’ SM use patterns and behaviors in clinical psychological science are discussed in the context of a pilot study using smartphone-based sensing with adolescents. The pilot study included 19 adolescents (Mean age = 15.84; 68% boys; 79% White) who installed a passive monitoring application (AWARE) on their phones for 31 (SD = 5.6) days. Descriptive data of sensing acceptability and feasibility are presented based on participant ratings and data yield ratio of usable data (74.18%). Sensing yielded 10,038 hourly observations collected from the ‘application foreground’ sensor across all participants from social media apps, and a total of 645 applications used. Categorization of SM apps were coded (kappa >.90) into ‘social networking’ (N = 20 apps) and ‘broader SM’ (N = 41) and compared to both Play Store-defined SM apps (N = 26) and popular SM apps based on Common Sense Media Survey (N = 9). Descriptive data on extracted behavioral features (duration, checking) from SM use categories (binned hourly and daily) are presented. Challenges, opportunities, and future directions of sensing methods for SM use are discussed to inform our understanding of its impacts on mental health and to improve the rigor of SM research in clinical psychological science.

Win Your Race Goal: A Generalized Approach to Prediction of Running Performance.

We introduce a novel approach for predicting running performance, designed to apply across a wide range of race distances (from marathons to ultras), elevation gains, and runner types (front-pack to back of the pack). To achieve this, the entire running logs of 15 runners, encompassing a total of 15,686 runs, were analyzed using two approaches: (1) regression and (2) time series regression (TSR). First, the prediction accuracy of a long short-term memory (LSTM) network was compared using both approaches. The regression approach demonstrated superior performance, achieving an accuracy of 89.13% in contrast, the TSR approach reached an accuracy of 85.21%. Both methods were evaluated using a test dataset that included the last 15 runs from each running log. Secondly, the performance of the LSTM model was compared against two benchmark models: Riegel formula and UltraSignup formula for a total of 60 races. The Riegel formula achieves an accuracy of 80%, UltraSignup 87.5%, and the LSTM model exhibits 90.4% accuracy. This work holds potential for integration into popular running apps and wearables, offering runners data-driven insights during their race preparations.

An empirical study of AI techniques in mobile applications

The integration of artificial intelligence (AI) into mobile applications has significantly transformed various domains, enhancing user experiences and providing personalized services through advanced machine learning (ML) and deep learning (DL) technologies. AI-driven mobile apps typically refer to applications that leverage ML/DL technologies to perform key tasks such as image recognition and natural language processing. Despite existing research exploring how mobile apps exploit AI techniques, they have the following main limitations: (1) Most existing studies focus on DL-based apps, with limited research on ML-based apps. (2) Existing research typically focuses on investigating the apps and the technologies utilized in the apps, lacking user-level analysis. (3) The number of apps studied is limited, with only 1,000 to 2,000 ML/DL apps identified after filtering. To fill the gap, in this paper, we conducted the most extensive empirical study on AI applications, exploring on-device ML apps, on-device DL apps, and AI service-supported (cloud-based) apps. Our study encompasses 56,682 real-world AI applications, focusing on three crucial perspectives: (1) Application analysis, where we analyze the popularity of AI apps and investigate the update states of AI apps; (2) Framework and model analysis, where we analyze AI framework usage and AI model protection; (3) User analysis, where we examine user privacy protection and user review attitudes. Our study has strong implications for AI app developers, users, and AI R&D. On one hand, our findings highlight the growing trend of AI integration in mobile applications, demonstrating the widespread adoption of various AI frameworks and models. On the other hand, our findings emphasize the need for robust model protection to enhance app security. Additionally, our study highlights the importance of user privacy and presents user attitudes towards the AI technologies utilized in current AI apps. We provide our AI app dataset (currently the most extensive AI app dataset) as an open-source resource for future research on AI technologies utilized in mobile applications.

Nutrition-Related Mobile Apps in the Spanish App Stores: Quality and Content Analysis.

Mobile apps represent accessible and cost-effective tools to improve nutrition and prevent chronic diseases. However, most of these apps have been characterized as having limited functionality, raising concerns about their effectiveness, acceptability, and efficacy. The aims of the study were to assess the quality of popular nutrition-related app platforms in Spain and to describe their characteristics and functionalities. We screened apps providing information on dietary advice, food advice, and nutritional content in the Apple App Store and Google Play Store in Spain from March 2 to March 16, 2024. Apps with a star rating of ≥4 (of 5 stars), those available in Spanish, those that were free of charge, those last updated after January 2022, those with >500 reviews, and those with >500,000 downloads were included. The quality of apps was assessed using the user version of the Mobile App Rating Scale (uMARS). General characteristics and nutritional, health, and market-related functionalities of the nutrition-related apps were described. Correlations among total and uMARS sections, star ratings, and number of reviews and downloads were evaluated. Among the 1460 apps identified in the search, 42 apps met the criteria. The majority of these (n=20, 48%) aimed at recording and analyzing food intake, followed by those providing nutritional plans or diets (n=9, 21%), advising on healthy habits (n=7, 17%), and offering recipes (n=6, 14%). The most prevalent nutritional functionalities offered were recording and monitoring body measurements (n=30, 71%), food tracking (n=26, 62%), and dietary analysis (n=25, 60%), whereas nutrition education was less common (n=16, 38%). Among market-related functionalities, advertisements were the most common among the study apps (n=30, 71%), followed by the option of sharing on social media (n=29, 69%) and customizable reminders (n=26, 62%). Sharing the recorded information in the app with health professionals was infrequent (n=1, 2%). The mean (SD) total uMARS score (maximum 5 points) was 3.78 (0.35), while the mean (SD) uMARS scores for functionality, aesthetics, engagement, and information were 4.21 (0.38), 3.94 (0.54), 3.51 (0.46), and 3.48 (0.44), respectively. Lower mean scores were observed for the subjective quality (mean 2.65, SD 0.56) and perceived impact (mean 3.06, SD 0.67). Moderate to strong positive significant correlations were mostly observed between total uMARS and section-specific uMARS scores, while the correlations between the uMARS section scores were mostly moderate positive. Total uMARS scores were very weakly correlated with user rating, number of reviews, and number of downloads. The quality of popular nutrition-related app platforms in Spain was acceptable, with observed remarkable differences between sections. The majority of the apps were appealing due to their user-friendly interfaces. Only a few apps, however, provided dietary structure analysis or nutritional education. Further research is needed to assess the long-term impact of these apps on users.

The Medium is the Message: How Secure Messaging Apps Leak Sensitive Data to Push Notification Services

Like most modern software, secure messaging apps rely on thirdparty components to implement important app functionality. Although this practice reduces engineering costs, it also introduces the risk of inadvertent privacy breaches due to misconfiguration errors or incomplete documentation. Our research investigated secure messaging apps' usage of Google's Firebase Cloud Messaging (FCM) service to send push notifications to Android devices. We analyzed 21 popular secure messaging apps from the Google Play Store to determine what personal information these apps leak in the payload of push notifications sent via FCM. Of these apps, 11 leaked metadata, including user identifiers (10 apps), sender or recipient names (7 apps), and phone numbers (2 apps), while 4 apps leaked the actual message content. Furthermore, none of the data we observed being leaked to FCM was specifically disclosed in those apps' privacy disclosures. We also found several apps employing strategies to mitigate this privacy leakage to FCM, with varying levels of success. Of the strategies we identified, none appeared to be common, shared, or well-supported. We argue that this is fundamentally an economics problem: incentives need to be correctly aligned to motivate platforms and SDK providers to make their systems secure and private by default.

Measuring Conditional Anonymity - A Global Study

The realm of digital health is experiencing a global surge, with mobile applications extending their reach into various facets of daily life. From tracking daily eating habits and vital functions to monitoring sleep patterns and even the menstrual cycle, these apps have become ubiquitous in their pursuit of comprehensive health insights. Many of these apps collect sensitive data and promise users to protect their privacy - often through pseudonymization. We analyze the real anonymity that users can expect by this approach and report on our findings. More concretely: We introduce the notion of conditional anonymity sets derived from statistical properties of the population; We measure anonymity sets for two real-world applications and present overarching findings from 39 countries; We develop a graphical tool for people to explore their own anonymity set. One of our case studies is a popular app for tracking the menstruation cycle. Our findings for this app show that, despite their promise to protect privacy, the collected data can be used to identify users up to groups of 5 people in 97% of all the US counties, allowing the de-anonymization of the individuals. Given that the US Supreme Court recently overturned abortion rights, the possibility of determining individuals is a calamity.

A comparative analysis of game experience in treadmill running applications

Despite the growing popularity of treadmill running apps, comparative analyses between exergames and traditional running simulators in terms of user experience are scarce. This study aims to fill this gap by examining user experiences across three different treadmill running apps: Zwift (a running simulator), Ring Fit Adventure (a Nintendo Switch game), and Fitoon (a mobile-specific treadmill game). Using unique interaction mechanisms-pedometer, leg-mounted controller with Ring-Con, and smartwatch, respectively, each app is evaluated through the In-Game version of the Game Experience Questionnaire. Results indicate that exergames, particularly Fitoon, significantly improve Competence, Immersion, and game flow compared to traditional treadmill running activities. However, contrary to expectations, the use of a smartwatch for interaction did not significantly decrease levels of Negative Affect. This finding underscores the complexity of the relationship between interaction mechanisms and their impact on user emotions within treadmill running applications. By providing a comprehensive analysis of user experiences, this study provides insight into the potential benefits and limitations of different interaction systems and platforms, providing valuable information for the development of more emotionally engaging and rewarding treadmill running apps. crucial.

Effectiveness of an mHealth App That Uses Financial Incentives and Gamification to Promote Health Behavior Change in Adolescents and Caregivers: Protocol for a Clinic-Based Randomized Controlled Trial.

Adolescent and adult obesity continues to be a public health epidemic in the United States. Despite the popularity of mHealth apps with gamification among adolescents, there are insufficient studies to evaluate the efficacy of gamified mHealth apps and financial incentives to motivate sustained health behavior change in adolescents or their adult caregivers. This study aims to evaluate the effectiveness of gamification techniques and financial incentives used in the novel "CommitFit" mHealth app to motivate health behavior change and improve various mental and physical health metrics in adolescents and their caregivers. This study is a 3-month randomized controlled trial (RCT) with 30 adolescents (aged 13-15 years) and their adult caregivers (N=60). It evaluates "CommitFit," which uses gamification including points and leaderboards to motivate logging and achievement of self-selected health behavior goals (eg, more water, sleep, physical activity, fruits, or vegetables or fewer sugary beverages). The RCT had three arms, each with 10 dyads: (1) CommitFit-only users; (2) CommitFit$, where adolescents were paid US $0.05 for each point they earned; and (3) waitlist control. Intervention dyads used the app for 3 months and had the option to use it for the fourth month without prompts or extra financial incentives. User analytic software was used to evaluate the frequency of user logs and goal achievement. Monthly surveys evaluated self-reported change in the 5 CommitFit health behaviors. Changes in BMI and blood pressure were evaluated for all participants at 3 clinical visits. Mental health, gamification, and behavior economics surveys were completed during the clinical visits. Recruitment began in August 2023 and was completed in 10 weeks. The research team successfully recruited and enrolled 30 dyads. Researchers emailed and called 89 caregivers on a physician-approved adolescent patient list, a 33% recruitment rate. Data collection and analysis will be conducted in the spring and summer of 2024. The results of this study are anticipated to be published between late 2024 and early 2025. This RCT will expand knowledge of the effectiveness of gamification techniques, financial incentives, and mHealth apps to motivate sustained health behavior change among adolescents and caregivers. These results may offer new opportunities to caregivers, health insurers, health care systems, and clinicians to motivate health behavior change in adolescents and caregivers, with the ultimate goal of preventing or reducing obesity and obesity-related diseases. Additional gamification, mental health surveys, and app user analytics included in the study may provide further insight into the characteristics of adolescents or caregivers who would benefit the most from using a gamified mHealth app like CommitFit. DERR1-10.2196/63505.

Botnet Attack Identification and Mitigation condition Software-Defined Networks Utilizing CNN Algorithm

One new design that makes managing and communicating across large-scale networks easier and more flexible is software-defined networking, or SDN. It allows for the smooth and dynamic execution of complicated network choices via programmable and centralized interfaces. But SDN opens doors for people and companies to tailor network apps to their needs, allowing them to enhance services. On the other hand, it began to encounter a host of new privacy and security issues and brought the dangers of one point of failure all at once. In most cases, hackers use OpenFlow switches to conduct botnets or distributed Denial of Service (DDoS) assaults against the controller. Popular security apps that use deep learning (DL) to quickly identify and counteract attacks are on the rise. Here, we examine botnet-based DDoS attack detection using DL approaches in an SDN-supported context and demonstrate their performance. For the assessment, we utilize a dataset that we just created ourselves. In order to choose the most useful subset of characteristics, we used weighting of features and tuning techniques. Using both a synthetic dataset and actual testbed conditions, we validate the measurements or simulation results. The primary objective of this research is to identify botnet-based DDoS assaults using easily-obtained characteristics and data using a lightweight DL approach with baseline hyper-parameters. We found that the DL technique's performance is affected by the optimal subset of features, and that the accuracy of predictions of the same approach may be varied with a different collection of features. Lastly, our empirical findings show that the CNN approach works better than both the dataset and the actual testbed environments. With CNN, the detection rate for typical flows is 99% and for malicious flows it drops to 97%.

Acceptability of an automated directly observed therapy (DOT) application for PrEP adherence support among young men who have sex with men: a qualitative exploration

ABSTRACT Pre-exposure prophylaxis (PrEP) adherence remains a challenge among young men who have sex with men (MSM). We developed and tested a smartphone application (“app”), “DOT Diary”, which combines automated directly observed therapy (DOT) with information about PrEP protection levels, pill-taking reminders, a sexual behavior diary, and a PrEP dosing calendar. To contextualize trial results, we qualitatively explored participants’ app experiences. The trial enrolled 100 young MSM in San Francisco and Atlanta. Participants were randomized 2:1 to DOT Diary versus standard-of-care and followed for 24 weeks. Interviews were conducted with 24 intervention participants. Data were analyzed using a memo-writing approach. Most expressed overall satisfaction with the app (“it was good for its purpose”), despite concerns about technical glitches. The most popular app features were the monthly calendar showing days PrEP was taken and information about level of protection based on pills taken. The DOT component helped participants establish PrEP routines. The reminders were “annoying but effective” at motivating dosing. Opinions about the sexual behavior diary varied. Overall, DOT Diary was acceptable; participants were willing to use it daily to record pill-taking. Critical components included the information about PrEP protection levels and calendar, while others may be modified to improve future success. Trial registration: ClinicalTrials.gov identifier: NCT03771638.

Unveiling Shopee App Continuance Use: Insights from Malaysian Shoppers

The popularity and acceptance of mobile shopping apps have surged widely due to mobile technology advancement and the COVID-19 outbreak. Retaining customers would be challenging for all mobile shopping app platform operators, particularly in the post-Covid era. However, research on the continuance of mobile shopping apps must be addressed. Hence, the Shopee mobile shopping app was adopted as a case study to examine antecedent factors of continuance intention leading to its continuance use among consumers. This quantitative study was conducted in Klang Valley, Malaysia, and data was collected via a self-administered Google Form. The study included 386 respondents and was analysed using Partial Least Squares Structural Equation Model (PLS-SEM). Customer loyalty, the success quality factors of the information system, incentives, trust, and satisfaction positively affect Shopee's mobile shopping app's continuance intention. Furthermore, it is found that continuance intention does indeed have a crucial effect on continuance use. However, the study reported that perceived enjoyment did not influence continuance intention. The objective of this research is achieved as it manages to generate valuable knowledge on continuance intention and use. This paves the way for more effective development of sustainable user retention strategies.

How can we improve buyer experiences for the operations of e-grocery platforms? Towards a unified framework for digital service operations

Electronic grocery (e-grocery) retailing has increasingly gained prominence all over the world, with India being no exception. However, the Indian e-grocery industry still has room to grow. Existing studies have not attempted to focus holistically on factors related to service operations, such as customer service and digital platforms. We undertake two case studies, each having an exploratory and a confirmatory study, on India's two popular e-grocery apps. We mine 273,000 and 254,000 buyer reviews to identify factors affecting their experiences in the exploratory studies. We model these factors through the lens of customer-dominant logic (CDL), service-dominant logic (SDL), and signaling theory. We develop an overarching framework by mapping these factors to theoretical lexicons for the confirmatory studies. Multiple econometric analysis helps us to confirm that trust in the vendor, consumer learning, retailer visibility, pricing, diagnosticity of product description, third-party product assurances, navigability, website informativeness, and perceived information protection influence the buyer experience of both e-grocery platforms significantly. These insights can help managers take appropriate decisions to improve buyer experience and thus expand their customer base.

Time matters: a critical multimodal study of an English learning app for children in China

ABSTRACT Time and its management, as evident in the scheduling, sequencing and structuring of activities, is important in early childhood education (ECE). While ECE apps have been changing teaching and learning practices for over a decade, research has yet to examine how they represent and manage time. To address this gap, we analysed the representation and management of time in iHuman ABC, a popular English learning app for children in China. Adopting Djonov and Van Leeuwen's (2018) model for analysing semiotic software and Van Leeuwen’s (2008) critical discourse studies framework, we examined temporal features in the app's design and references to these features in promotional materials and user reviews. Findings reveal that iHuman ABC reflects an instructivist ideology which considerably limits children’s ability to organise their time and contradicts research-based recommendations for teaching English. This study contributes to critical multimodal studies of the potential of software to transform educational practices.

Evaluating the Quality and Comparative Validity of Manual Food Logging and Artificial Intelligence-Enabled Food Image Recognition in Apps for Nutrition Care.

For artificial intelligence (AI) to support nutrition care, high quality and accuracy of its features within smartphone applications (apps) are essential. This study evaluated popular apps' features, quality, behaviour change potential, and comparative validity of dietary assessment via manual logging and AI. The top 200 free and paid nutrition-related apps from Australia's Apple App and Google Play stores were screened (n = 800). Apps were assessed using MARS (quality) and ABACUS (behaviour change potential). Nutritional outputs from manual food logging and AI-enabled food-image recognition apps were compared with food records for Western, Asian, and Recommended diets. Among 18 apps, Noom scored highest on MARS (mean = 4.44) and ABACUS (21/21). From 16 manual food-logging apps, energy was overestimated for Western (mean: 1040 kJ) but underestimated for Asian (mean: -1520 kJ) diets. MyFitnessPal and Fastic had the highest accuracy (97% and 92%, respectively) out of seven AI-enabled food image recognition apps. Apps with more AI integration demonstrated better functionality, but automatic energy estimations from AI-enabled food image recognition were inaccurate. To enhance the integration of apps into nutrition care, collaborating with dietitians is essential for improving their credibility and comparative validity by expanding food databases. Moreover, training AI models are needed to improve AI-enabled food recognition, especially for mixed dishes and culturally diverse foods.

The Dark Side of Recommendation Algorithms in Chinese Mass Short Video Apps: Effect of Perceived Over-Recommendation on Users’ Cognitive Dissonance and Discontinuance Intention

In the digital age, human-algorithm interaction has become ubiquitous. The popularity of short-form video apps has led to greater access to and experience with algorithmic recommender systems, but it has also exposed their potential problems and dark sides. However, there is still limited knowledge about the dark side of over-recommendation from the perspective of user perception, especially regarding the negative impact on users’ psychology and behavior. Drawing on social cognitive theory, this study constructed a theoretical model to investigate how information environment factors influence users’ cognitive dissonance and discontinuance intention under the antecedent dimension of perceived over-recommendation. The model was tested using partial least squares structural equation modeling on 322 valid questionnaires from users of Chinse mass short video apps (eg, Douyin). The results showed that when users perceived over-recommendation, information narrowing, information redundancy, perceived overload, and privacy invasion significantly increased their cognitive dissonance, ultimately leading to discontinuance intention. Notably, cognitive dissonance fully mediated the relationship between the information environment and discontinuance intention, and self-efficacy did not play a significant moderating role. Additionally, the local path effects varied significantly across groups with different characteristics. To maintain the sustainability of short video platforms, it is crucial to explore moderate recommendation mechanisms. User-centered functionality improvements and differentiated behavioral interventions can help mitigate negative psychology and discontinuance intention among users.

Boosting Smoking Cessation Intervention Utilization in Chinese Healthcare Providers: A Randomized Controlled Trial of the 'WeChat WeQuit' Medical Education Program.

In China, standard smoking cessation practices are rarely used by healthcare service providers (HSPs). WeChat, a popular social media app, has been widely used in China. In this single-blind, randomized trial, undertaken in China with 8-week interventions and follow-up to 34 weeks, 1887 HSPs were randomly selected to the intervention (n=942) or control group (n=945) from Oct 2020 to Oct 2021. The intervention group received regular smoking cessation training program messages from the professional team for 8 weeks and followed for 34 weeks. The control group received thanks messages for 8 weeks, and follow-up to 34 weeks. Both groups received a hard copy of the manual after randomization. The primary outcome measure was the utilization rate of behavioral and pharmacotherapy interventions for smoking patients from 9 to 34 weeks. This trial is registered at ClinicalTrials.gov (number NCT03556774). HSPs in the intervention group demonstrated a better overall utilization rate of smoking cessation at 20-week follow-up compared to the control group (35.54% vs 31.41%, p=0.036). Additionally, both groups showed a significant increase in the adoption of various components of the 5A's model - including "Assess", "Assist: set a quit date", "Assist: recommend cessation program", "Assist: provide information", "Assist: recommend medication", and "Arrange" - at the 9-week follow-up relative to baseline. Notably, at the 20-week follow-up, the intervention group reported significantly enhanced utilization rates for all these components, except "Assist: set a quit date". The "Wechat Wequit" training program effectively enhanced smoking cessation intervention adoption among Chinese HSPs. 'WeChat WeQuit' training program was effective in increasing the provision of effective tobacco cessation interventions by Chinese-speaking HSPs to patients with cigarette smoking, which could provide valuable insights into bridging the gap between need and services for smoking cessation in China.