The availability of mobile money in Kenya has positively impacted commerce, financial transaction efforts, and the ability of individuals to receive and save their money. The addition of mobile loan applications provides access to loans without the hassle of going to a physical bank and, in some cases, completing paperwork. While it has its benefits, limited protection of user data has been a cause for concern. User complaints prompted a change in the mobile loan industry, requiring applications to be licensed and banning the use of specific permissions for Android versions of the apps placed in the Google Play Store. We investigate the impact of this change and explore ways to improve regulation by reviewing 30 licensed (n=15) and unlicensed (n=15) Kenyan-targeted digital lender apps. The results suggest that regulation has not yet had a significant impact on digital lender app development and thus encourages government-supported development guidelines and audits.
Sudden Infant Death Syndrome (SIDS) is a serious public health concern, representing one of the leading causes of mortality in infants during their first year of life. This study focuses on investigating the possible relationship between sleep apnea and SIDS, addressing factors such as birth weight, age, and gender of the neonates. Using a methodology of data mining and recurrent neural networks (RNNs), physiological data obtained during the first 12 hours of life were analyzed. Initial results showed limited correlations, but with deeper analysis, greater consistency and accuracy in apnea predictions were achieved, reaching high reliability in the RNN model. Collaboration with the Mexican Red Cross was crucial, providing valuable data and allowing for the reevaluation of neonatal monitoring protocols. This methodological approach not only enhances the detection and early prevention of SIDS based on sleep apnea but also suggests practical solutions to reduce the incidence of this tragic condition.
Nowadays, artificial intelligence (AI) is finding relevance in diverse facets of human existence. For example, AI has been applied in areas such as finance, energy, healthcare, transportation, education, agriculture, and entertainment, just to mention a few. However, technologies that employ AI algorithms such as machine learning algorithms and deep learning algorithms are susceptible to a new challenge known as adversarial machine learning (AML) attacks. An AML attack occurs when a malicious actor crafts adversarial examples (AEs) to introduce perturbation into the original input data of an AI-based classification model to cause the model to misclassify inputs. This paper reviewed how computer vision systems, audio systems, industrial systems, transportation systems, blockchain technologies, and cybersecurity systems may be susceptible to AML attacks. Special attention is given to the discussion on AML attacks on network intrusion detection systems (NIDSs) while also highlighting adversarial defense strategies. The paper concludes with a proposal for future research direction regarding AML attack mitigation.
This paper introduces work by the IEEE P7014.1 Working Group on the Recommended Practice for Ethical Considerations of Emulated Empathy in Partner-based General-Purpose Artificial Intelligence Systems. This paper briefly details the scope and parameters of the standard, why it matters, and key ethical problems found regarding use of modern AI systems that emulate empathy for human AI-partnering. Some of these problems are fairly obvious, and others are less so, but no less important. A few however require deeper consideration because, like many important ethical discussions, they do not have easy answers. One such question is when is deception in human-computer interaction acceptable, particularly where deception overlaps with animism and anthropomorphism and may be exacerbated by emulations of empathy? This paper lingers on this question, drawing on philosophical and ethical discussion about the nature of deception, contexts where it is acceptable and beneficial, and contexts where it is morally out of scope.
This paper examines the integration of large language models (LLMs) in the workplace using a hermeneutic phenomenological approach, highlighting the critical differences between human language use and LLM operational mechanisms. It challenges the uncritical adoption of LLMs, emphasizing their limitations in contextual understanding and intuitive decision-making. Advocating a reimagined workforce, the paper proposes viewing LLMs as "AI interns" and humans as "Master Orchestrators," focusing on enhancing human creativity and ingenuity to ensure AI supports rather than replaces human capabilities.
