Social media platforms like Instagram significantly influence tourists’ travel decisions by providing them with valuable insights, recommendations, authentic information, and points of interest. However, photos shared with location-specific hashtags, even those related to tourist attractions, do not always reflect the actual destination, creating challenges for potential visitors seeking accurate information. To assist tourists in finding pertinent tourism information for specific destinations, we propose VISTA: Visual Identification of Significant Travel Attractions. The proposed method employs deep learning and active learning techniques to automatically classify photos into: ‘Tourism-Related’ photos (i.e., photos related to tourism) and ‘Non-Tourism-Related’ photos (i.e., photos unrelated to tourism). To train our machine learning classifier, we created a dataset containing photos of the 10 most popular Israeli cities on Instagram. The classifier obtained an accuracy score of 0.965 and a weighted F1 score of 0.964. Evaluating our classifier’s global generalization on the InstaCities100K dataset, derived from InstaCities1M, yielded an accuracy score of 0.958 and a weighted F1 score of 0.959. The effectiveness of VISTA was demonstrated by comparing tourism-related and non-tourism-related photos in terms of photo proportion, user engagement, and object comparison. We found that most photos published on Instagram associated with cities are irrelevant to tourists and that tourism-related photos received more likes than non-tourism-related photos. Finally, there was a low overlap between objects in the two photo collections. Based on these results, we conclude that VISTA can help tourists tackle the problem of finding relevant tourism-related photos among the high volume of photos available on Instagram.

Background: Searching for web-based health-related information is frequently performed by the public and may affect public behavior regarding health decision-making. Particularly, it may result in anxiety, erroneous, and harmful self-diagnosis. Most searched health-related topics are cancer, cardiovascular diseases, and infectious diseases. A health-related web-based search may result in either formal or informal medical website, both of which may evoke feelings of fear and negativity. Objective: Our study aimed to assess whether there is a difference in fear and negativity levels between information appearing on formal and informal health-related websites. Methods: A web search was performed to retrieve the contents of websites containing symptoms of selected diseases, using selected common symptoms. Retrieved websites were classified into formal and informal websites. Fear and negativity of each content were evaluated using 3 transformer models. A fourth transformer model was fine-tuned using an existing emotion data set obtained from a web-based health community. For formal and informal websites, fear and negativity levels were aggregated. t tests were conducted to evaluate the differences in fear and negativity levels between formal and informal websites. Results: In this study, unique websites (N=1448) were collected, of which 534 were considered formal and 914 were considered informal. There were 1820 result pages from formal websites and 1494 result pages from informal websites. According to our findings, fear levels were statistically higher (t2753=3.331; P<.001) on formal websites (mean 0.388, SD 0.177) than on informal websites (mean 0.366, SD 0.168). The results also show that the level of negativity was statistically higher (t2753=2.726; P=.006) on formal websites (mean 0.657, SD 0.211) than on informal websites (mean 0.636, SD 0.201). Conclusions: Positive texts may increase the credibility of formal health websites and increase their usage by the general public and the public’s compliance to the recommendations. Increasing the usage of natural language processing tools before publishing health-related information to achieve a more positive and less stressful text to be disseminated to the public is recommended.

Online social networks (OSNs) are fertile ground for information sharing and public relationships. However, the uncontrolled dissemination of information poses a significant risk of the inadvertent disclosure of sensitive information. This poses a notable challenge to the information security of many organizations. Improving organizations’ ability to automatically identify data leaked within image-based content requires specialized techniques. In contrast to traditional vision-based tasks, detecting data leaked within images presents a unique challenge due to the context-dependent nature and sparsity of the target objects, as well as the possibility that these objects may appear in an image inadvertently as background or small elements rather than as the central focus of the image. In this paper, we investigated the ability of multiple state-of-the-art deep learning methods to detect censored objects in an image. We conducted a case study utilizing Instagram images published by members of a large organization. Six types of objects that were not intended for public exposure were detected with an average accuracy of 0.9454 and an average macro F1-score of 0.658. A further analysis of relevant OSN images revealed that many contained confidential information, exposing the organization and its members to security risks.

Celiac disease and gluten sensitivity affect a significant portion of the population and require adherence to a gluten-free diet. Dining in social settings, such as family events, workplace gatherings, or restaurants, makes it difficult to ensure that certain foods are gluten-free. Despite the availability of portable gluten testing devices, these instruments have high costs, disposable capsules, depend on user preparation and technique, and cannot analyze an entire meal or detect gluten levels below the legal thresholds, potentially leading to inaccurate results. In this study, we propose RGB (Recognition of Gluten in Bread), a novel deep learning-based method for automatically detecting gluten in bread images. RGB is a decision-support tool to help individuals with celiac disease make informed dietary choices. To develop this method, we curated and annotated three unique datasets of bread images collected from Pinterest, Instagram, and a custom dataset containing information about flour types. Fine-tuning pre-trained convolutional neural networks (CNNs) on the Pinterest dataset, our best-performing model, ResNet50V2, achieved 77% accuracy and recall. Transfer learning was subsequently applied to adapt the model to the Instagram dataset, resulting in 78% accuracy and 77% recall. Finally, further fine-tuning the model on a significantly different dataset, the custom bread dataset, significantly improved the performance, achieving an accuracy of 86%, precision of 87%, recall of 86%, and F1-score of 86%. Our analysis further revealed that the model performed better on gluten-free flours, achieving higher accuracy scores for these types. This study demonstrates the feasibility of image-based gluten detection in bread and highlights its potential to provide a cost-effective non-invasive alternative to traditional testing methods by allowing individuals with celiac disease to receive immediate feedback on potential gluten content in their meals through simple food photography.

Trigger warnings are placed at the beginning of potentially distressing content to provide individuals with the opportunity to avoid the content before exposure. Social media platforms use artificial intelligence to add automatic trigger warnings to certain images and videos, but are less commonly applied to textual content. This leaves the responsibility of adding trigger warnings to the authors, and a failure to do so may expose vulnerable users to sensitive or upsetting content. Due to limited research attention, there is a lack of understanding concerning what content is or is not considered triggering by social media users. To address this gap, we examine the use of trigger warnings in tweets on X, previously known as Twitter. We used a large language model (LLM) for zero-shot learning to identify the types of trigger warnings (e.g., violence, abuse) used in tweets and their prevalence. Additionally, we employed sentiment and emotion analysis to examine each trigger warning category, aiming to identify prevalent emotions and overall sentiment. Two datasets were collected: 48,168 tweets with explicit trigger warnings and 4,980,466 tweets with potentially triggering content. The analysis of the smaller dataset indicates that users have applied trigger warnings more frequently over the years and are applying them to a broader range of content categories than they did in the past. These findings may reflect users’ growing interest in creating a safe space and a supportive online community that is aware of diverse sensitivities among users. Despite these findings, our analysis of the larger dataset confirms a lack of trigger warnings in most potentially triggering content.

Background: Social media is a vital tool for health organizations, enabling them to share evidence-based information, educate the public, correct misinformation, and support a more informed and healthier society. Objective: This study aimed to categorize health organizations’ content on social media into topics; examine public engagement, sentiment, and emotional responses to these topics; and identify gaps in fear between health organizations’ messages and the public response. Methods: Real data were collected from the official Instagram accounts of health organizations worldwide. The BERTopic algorithm for topic modeling was used to categorize health organizations’ posts into distinct topics. For each identified topic, we analyzed the engagement metrics (number of comments and likes) of posts categorized under the same topic, calculating the average engagement received. We examined the sentiment and emotional content of both posts and responses within the same topic, providing insights into the distributions of sentiment and emotions for each topic. Special attention was given to identifying emotions, such as fear, expressed in the posts and responses. In addition, a linguistic analysis and an analysis of sentiments and emotions over time were conducted. Results: A total of 6082 posts and 82,982 comments were collected from the official Instagram accounts of 8 health organizations. The study revealed that topics related to COVID-19, vaccines, and humanitarian crises (such as the Ukraine conflict and the war in Gaza) generated the highest engagement. Our sentiment analysis of the responses to health organizations’ posts showed that topics related to vaccines and monkeypox generated the highest percentage of negative responses. Fear was the dominant emotion expressed in the posts’ text, while the public’s responses showed more varied emotions, with anger notably high in discussions around vaccines. Gaps were observed between the level of fear conveyed in posts published by health organizations and in the fear conveyed in the public’s responses to such posts, especially regarding mask wearing during COVID-19 and the influenza vaccine. Conclusions: This study underscores the importance of transparent communication that considers the emotional and sentiment-driven responses of the public on social media, particularly regarding vaccines. Understanding the psychological and social dynamics associated with public interaction with health information online can help health organizations achieve public health goals, fostering trust, countering misinformation, and promoting informed health behavior.