UMT Artificial Intelligence Review

AI and BI Synergy: A New Frontier in Business and Economics

Muhammad Ahmad Mazher — 2024-05-15

The convergence of Artificial Intelligence (AI) and Business Intelligence (BI) is creating a new era of transformative change in the realms of business and economics. The current study aimed to combine AI with standard BI approaches in order to achieve results, such as enhanced decision-making, optimized operations, and a competitive edge. The merger of BI and AI helps organizations and businesses derive deeper insights while using big datasets, which may cause improvements in predictive skills, personalized customer experiences, and ultimately optimized allocation of available resources. Moreover, this study also examined the prior literature and founded on observations, investigated the interaction of BI and AI with economics and areas pertaining to business study areas. After establishing and executing prior literature reviews and analytical structures, we concluded that the AI and BI combination is necessary and meaningful for the analysis of economic issues as well as business problems. In the current era, AI and BI interventions can generate complex challenges, daily threats, and innovative ideas. The study revealed a positive and significant influence of BI and AI's positive and significant influence in resolve economic and business challenges and making decision-making an efficiently.

Examining the Perceived Performance of Artificial Intelligence on the Behavioral Front

Saman Javed — 2024-05-15

Artificial intelligence (AI) and its associated technologies have experienced rapid advancements, especially in the 21^st century. While the proficiency of knowledge-based AI is well-established, behavior-based AI still faces significant challenges. There exists uncertainty about the effectiveness of AI systems in performing behavioral roles, that typically belong to human beings. Based on AI-driven social theory, this research argues that the development of AI systems is closely intertwined with social facets. Since the foundation of all AI technologies is ingrained in anthropocentrism, the study inquires about the confidence of users/respondents in AI’s ability to assume behavioral roles. For empirical analysis, data was collected from 120 university students. Rudimentary scales were designed to gauge the influence of AI across eight behavioral parameters, namely sentience, personality, leadership, ethics, decision-making, power, conflict management, and emotions. Descriptive data analysis revealed somewhat vacillating results. Among the eight behavioral parameters, respondents showed high confidence in AI’s decision-making capabilities. Moreover, the results revealed respondents’ moderate confidence in AI’s ability to exercise power and manage conflicts. Conversely, confidence in AI’s emotional prowess is found to be relatively low. It is further found that females believe more in the prospect of AI sentience relative to males. No significant difference between male and female perceptions was found for the rest of the parameters. The study's indeterminate findings concluded that users are confident, as well as ambivalent of behavioral AI’s perceived performance. The middle-of-the-road results suggested skepticism around AI’s behavioral capabilities.

Detection and Prevention of DNS Tunneling Attacks: Exploring Technologies and Methodologies

Usman Inayat — 2024-05-15

DNS tunneling attack is one of the most common and ignored attacks that the current systems are vulnerable to. This study examines the functionality of DNS in terms of DNS hierarchy and the ways through which intruder creates a tunnel. The research used both rule-based and model-based technology tools alongwith other detection-based technologies, namely signature-based and threshold-based technologies. The graphical representation of the tunnel detection technology has been shown to better understand the systematic working of DNS. Based on the review of previous research methodologies, the current research analysed methods for the detection and prevention of DNS tunneling, which includes a location-based model using GPS and observing data packet sizes.

Machine Learning-Based Suicide Risk Assessment and Intervention Strategies for Depression

Muhammad Yousif — 2024-05-15

Suicide is a global issue, primarily caused by depression. Over the past three decades, the World Health Organization reports that a considerable number of people have died by suicide. This study uses machine-learning models like Naive Bayes and logistic regression, to predict suicide risk using a dataset of social media posts. Previous research has used SVM and random forest, but deep learning techniques could improve accuracy by analyzing visual and auditory data. This would simplify mental health professionals' work and move away from traditional methods. In today’s digital world, leveraging digital tools can make significant progress in suicide prevention and mental health support. Moreover, future developments may include refined clinical reports with human experts, providing researchers with more effective tools for improving mental health outcomes