Tag: Climate Change Disinformation

Abstract

As entering the third decade of the twenty-first century, we face two emerging issues: the aging global population and the widespread applications of artificial intelligence. This article presents a framework for discussing how population aging and artificial intelligence directly affect households and living arrangements and sustainable development. Population aging and artificial intelligence may also impact sustainable development indirectly through households and living arrangements. Artificial intelligence offers a promising future with the potential to boost productivity, reduce anthropogenic impact and improve society. It also creates new jobs for developing, manufacturing and maintaining artificial intelligence products and services. However, it is crucial to consider potential concerns with societal impacts, increased consumption of non-renewable sources and power, unemployment and ethical issues. We conclude with policy recommendations for government implementation and academic research, emphasizing the need for careful consideration and planning in integrating artificial intelligence into sustainable development, as this will be crucial for the future of our society. Exploring the crucial interplay of population aging and artificial intelligence on households, living arrangements and sustainable development will be a vital topic for researchers, policymakers and academics in sociology, demography, economics, education and environmental and public health.

Abstract

The urgent need for a net-zero future necessitates a fundamental shift in the energy sector, with road transportation responsible for a substantial 37% of global energy-related CO₂ emissions in 2021, emerging as a pivotal focal point in the battle against climate change. Energy consumption in the road sector is expected to surge by 1.26% with a 1% growth in urbanization, concentrated mainly in Asia and Africa by the mid-2030s. Therefore, addressing emissions from the transportation industry is paramount. Electric vehicles (EVs), coupled with a transition to renewable energy, offer a sustainable solution, yet their market share remains at a modest 10% globally and in Asia. With numerous nations committed to achieving net-zero emissions, EV adoption is on the rise, particularly in developing regions with high urbanization and Greenhouse Gas (GHG) emissions. Governments worldwide have initiated policies that provide incentives to promote EVs, but challenges like patent declines and EV battery disposal concerns persist. In this paper, we make an integrative critical review of the existing literature, conduct a SWOT analysis of EVs, and address crucial factors influencing their adoption, thereby contributing to the goal of a more sustainable future in road transportation.

Abstract

Topic modelling (TM) is a significant natural language processing (NLP) task and is becoming more popular, especially, in the context of literature synthesis and analysis. Despite the growing volume of studies on the use of and versatility of TM, the knowledge of TM development, especially from the perspective of bibliometrics analysis is limited. To this end, this study evaluated TM research using two techniques namely, bibliometrics analysis and TM itself to provide the current status and the pathway for future studies in the TM field. For this purpose, this study used 16,941 documents collected from Scopus database from 2004 to 2023. Results indicate that the publications on TM have increased over the years, however, the citation impact has declined. Furthermore, the scientific production on TM is concentrated in two countries namely, China and the USA. Our findings showed there are several applications of TM that are understudied, for example, TM for image segmentation and classification. This paper highlighted the future research directions, most importantly, calls for increased multidisciplinary research approaches to fully deploy TM algorithms optimally and thus, increase usage in non-computer science subject areas.

Abstract

Malaria remains a significant global health challenge, with nearly half of the world's population at risk of infection. In 2022 alone, malaria claimed approximately 608,000 lives, with 76% of these fatalities occurring in children under the age of five, underscoring the disease’s disproportionate impact on vulnerable populations. Africa bears the highest burden, accounting for 94% of global malaria cases. For over 60 years, the development of a malaria vaccine has been a critical objective for scientists and governments, with substantial efforts directed toward this goal. Recent progress has led to the approval of the first malaria vaccines, RTS,S/AS01 (Mosquirix^®) and the R21/Matrix-M vaccine. Inspired by the promise of these vaccines, the global malaria community has renewed its focus on malaria eradication, 50 years after flawed earlier eradication efforts in the mid-twentieth century. Since the World Health Organization’s endorsement of RTS,S in 2021 and R21 in 2023, several African countries, beginning with Cameroon, have integrated these vaccines into routine immunization programmes. This review examines the role of routine malaria vaccination in Africa as a key strategy toward malaria elimination, explores challenges and solutions for widespread vaccine implementation, and discusses future directions in the ongoing fight to eliminate malaria on the continent.

Abstract

Fake news poses a significant threat to public opinion and social stability in modern society. This study presents a comparative evaluation of BERT-like encoder-only models and autoregressive decoder-only large language models (LLMs) for fake news detection. We introduce a dataset of news articles labeled with GPT-4 assistance (an AI-labeling method) and verified by human experts to ensure reliability. Both BERT-like encoder-only models and LLMs were fine-tuned on this dataset. Additionally, we developed an instruction-tuned LLM approach with majority voting during inference for label generation. Our analysis reveals that BERT-like models generally outperform LLMs in classification tasks, while LLMs demonstrate superior robustness against text perturbations. Compared to weak labels (distant supervision) data, the results show that AI labels with human supervision achieve better classification results. This study highlights the effectiveness of combining AI-based annotation with human oversight and demonstrates the performance of different families of machine learning models for fake news detection.

Abstract

Evidence points to a mental health crisis among young people. To address their unmet mental health needs, students have turned to social media and other online platforms to seek mental help and advocate for better mental health. According to resource theory, tangible (e.g., social media and other digital technologies) and intangible resources (e.g., mental health literacy, social media competence) may bolster young people’s engagement in mental health advocacy. Drawing from a sample of 157 Filipino senior high school students from a private university, this convergent mixed methods study aims to (1) describe online mental health advocacy engagement; and (2) examine resources related to their engagement. Quantitative (i.e., scales) and qualitative data (i.e., open-ended questions) were collected via an online survey. Findings suggest that latent engagement (i.e., mental health information consumption) is the more commonly practiced online mental health advocacy engagement, followed by follower and expressive engagement (i.e., sharing mental health-related posts). Another identified form of online advocacy engagement is the provision of mental help via online platforms. Moreover, two major resources for online advocacy engagement surfaced from the integrated findings: digital affordances, which include social media presence and other related digital tools, and the “self,” which encompasses the youth advocate’s mental health literacy, social media competence, assessment of health (mis)information, and well-being. Insights from this study can assist mental health advocacy groups in designing and implementing initiatives to increase the participation of youth advocates.

Abstract

Situated affectivity investigates how natural, material, and social environmental structures, so-called ‘scaffolds,’ influence our affective life. Initially, the debate focused on user-resource-interactions, i.e., on cases where individuals (‘users’) actively structure the environment (‘resource’) in beneficial ways, setting up scaffolds that allow them to solve routine problems, modify their means of coping with challenges, or avail themselves of new affective competences. More recently, cases of mind invasion have captured philosophers’ attention where the ways others structure the environment affect, or invade, people’s minds, typically without their awareness and with harmful consequences. This paper contributes to recent discussions about the variety of phenomena that can count as ‘scaffolded affectivity’ in general and ‘scaffolded affective harm’ in particular. It also addresses the emerging question of how harmful affective scaffolds can come to have a grip on people’s minds, despite their detrimental consequences. We first disentangle some misconceptions and illustrate how diverse (harmful) affective scaffolds can be. In contrast to recent approaches that have characterized scaffolds in largely descriptive terms, we then identify factors that can help explain why a given scaffold is effective in modifying people’s minds. We also try to shed light on why some agents and some social structures are especially likely to experience or cause scaffolded affective harm, respectively, by arguing that user-resource-interactions and mind invasions are not independent, but intimately intertwined and mutually reinforcing, especially in the digital domain. We conclude with a speculative suggestion for further research.

Abstract

Large Language Models (LLMs) gained notable popularity in academia and industry. It has unprecedented features and performance in many applications. LLMs are revolutionizing the AI industry by enabling them to offer sophisticated and human-like natural language processing capabilities. LLMs empower applications to understand, interpret, and generate human-like text at a level that was previously challenging to achieve. On this, numerous LLM providers including Open AI, Google AI, Meta AI, and Microsoft have commenced to offer LLM services to their customers. The integration of LLMs into healthcare represents a significant advancement in the delivery of medical care. Alas, the diversity of LLMs presents a challenge for healthcare providers in determining which of these LLMs is the most appropriate option that meets the user’s requirements. This study aims to develop a LLM evaluation criteria framework that help healthcare providers to select the best LLM that meet their requirements. To do so, 38 experts in the domain of AI were interviewed and the fuzzy analytical hierarchy process (FAHP) is applied to compute the weight of these criteria. The results identified 9 evaluation criteria with 12 sub-criteria along with their specific metrics as the most critical criteria in evaluating and selecting LLMs in healthcare domain. The analysis results show that LLM evaluation criteria are ranked in descending order of importance, with assigned weights as follows: reliability (0.200), robustness (0.171), bias and fairness (0.126), availability (0.121), performance (0.092), usability (0.090), resilience (0.089), predictability (0.063), and cost (0.047). As a result, the proposed framework can provide useful feedback to healthcare providers by allowing them to select the best LLM that meets their requirements. Also, it can help providers by allowing them to satisfy their Service Level Agreement (SLA) and improve their quality of service.