Adverse Drug Events (ADEs) are a major healthcare issue in the United States, contributing to millions of outpatient and emergency department visits and ranking as the fourth leading cause of death. While many ADEs are identified post-market, improved detection methods are crucial for enhancing patient safety. This study explores the application of large language models (LLMs) to the n2c2 task for ADE detection, evaluating optimal prompting techniques without requiring ADE-specific training data. Results indicate that an entity-only extraction approach outperforms the inline method, offering higher precision, recall, and token efficiency. This study highlights the potential of LLMs for accurate ADE detection in clinical text, improving performance while maintaining model efficiency.

The intersection of pervasive technology and verbal communication has resulted in the creation of Automatic Speech Recognition Systems (ASRs), which automate the conversion of spontaneous speech into texts. ASR enables human-computer interactions through speech and is rapidly integrated into our daily lives. However, the research studies on current ASR technologies have reported unfulfilled social inclusivity and accentuated biases and stereotypes towards minorities. In this work, we provide a review of examples and evidence to demonstrate preexisting sexist behavior in ASR systems through a systematic review of research literature over the past five years. For each article, we also provide the ASR technology used, highlight specific instances of reported bias, discuss the impact of this bias on the female community, and suggest possible methods of mitigation. We believe this paper will provide insights into the harm that unchecked AI-powered technologies can have on a community by contributing to the growing body of research on this topic and underscoring the need for technological inclusivity for all demographics, especially women.

While humanity prepares for a post-pandemic world and a return to normality through worldwide vaccination campaigns, each country experienced different levels of impact based on natural, political, regulatory, and socio-economic factors. To prepare for a possible future with COVID-19 and similar outbreaks, it is imperative to understand how each of these factors impacted spread and mortality. We train and tune two decision tree regression models to predict COVID-related cases and deaths using a multitude of features. Our findings suggest that, at the country-level, GDP per capita and comorbidity mortality rate are best predictors for both outcomes. Furthermore, latitude and smoking prevalence are also significantly related to COVID-related spread and mortality.