Artificial intelligence (AI) and large language models are transforming assessment in computer science (CS) education, challenging academic integrity and knowledge retention. This article introduces a diagnostic framework and evaluates evidence-informed techniques, offering CS instructors a structured, practice-based guide for maintaining authentic learning in the age of AI.
Generative artificial intelligence (GAI) and emerging agentic AI systems are transforming the core functions of higher education, from course design and assessment to academic governance and IT strategy. This article proposes a four-layered framework that aligns policy, pedagogy, assessment, and capacity-building to help institutions move from ad hoc reactions to coordinated, responsible academic innovation. The framework translates high-level AI ethics and academic integrity principles into actionable steps, including policy guardrails, AI-aware course and assessment design, oversight of agentic AI systems, and sustainable capacity building for faculty and students. Grounded in interdisciplinary research, institutional case studies, and policy guidance from international organizations, this article illustrates how each layer can be implemented in both low-resource and more mature environments. The goal is to provide a practical, mission-aligned roadmap for integrating generative and agentic AI in ways that protect academic standards, support equity, and enable universities to shape rather than react to the AI transition.
This study explores the use of drones to capture images of natural environments like ecological conservation areas and national parks. The images are analyzed to identify harmful plants and species that need protection, with the results sent to relevant personnel via an app. We integrate 5G technology, unmanned aerial vehicles (UAVs) (drones), and artificial intelligence [You Only Look Once (YOLO) recognition system] to enhance image quality and recognition accuracy while ensuring low-latency communication. By upgrading from YOLO-V5 to YOLO-V8, we achieve better throughput and efficiency, particularly for identifying multiple plant species. The drones’ wide flight and shooting angles improve inspection efficiency. This system not only detects harmful plants but also recognizes rare herbs and conservation plants, enabling better management of natural environments. Our app enhances user experience and aligns with six United Nations Sustainable Development Goals, promoting ecological balance and fostering conservation awareness for sustainable development.
The use of large language models (LLMs) is being introduced into requirements, code generation, testing, maintenance, and documentation processes, but most IT organizations have yet to establish a practical and evidence-based methodology regarding when these tools are value added, when they become risky, and how to regulate their usage. The article is a synthesis of recent empirical research, surveys of developers, and guidance on the use of LLMs in software engineering and translates that information into a playbook of guidance that can be applied by practitioners. The primary contribution of the article is a staged adoption framework, which includes explore, pilot, and scale, supported with lightweight survey templates, small-task assessment designs, and accept/edit/reject logging practices that organizations can adopt to produce their own context-specific evidence. The objective is to facilitate disciplined, open-minded adoption of LLMs in actual software engineering environments.
Artificial intelligence (AI) transforms organizational operations, demanding responsible AI deployment workforce capabilities. AI literacy—combining technical fluency with ethical and societal awareness—is emerging as a critical organizational capability in the digital workplace. This article presents AI in Play, a scenario-based serious game designed to build responsible AI adoption skills among IT professionals and university students. The game was developed by one of the authors and integrates FATE principles—fairness, accountability, transparency, and ethics—into decision-making under uncertainty, simulating enterprise constraints such as data readiness, governance maturity, and stakeholder alignment. We explain the rationale behind each mechanic, map mechanics to targeted competencies, and report four exploratory demonstration sessions with 32 enrolled participants. Analytical autoethnography documents design decisions and reflective iteration; systematic evaluation of effectiveness is planned as future work. We believe serious-game pedagogy offers institutions a practical, human-centered approach to preparing their workforce for responsible AI adoption.
As the financial services sector transitions from predictive artificial intelligence to autonomous, agentic systems, traditional oversight frameworks are failing to keep pace. This article identifies the “velocity trap,” a structural disconnect where algorithmic execution speeds outrun the latency of manual compliance review. By introducing the “automation mandate,” this research proposes a shift toward “living compliance” through a dedicated hardware/logic architecture (U.K. Patent Class 14-02). This system replaces probabilistic monitoring with deterministic enforcement, utilizing symbolic execution to provide a “forensic reasoning trace.” This architecture elevates the human role from a manual reviewer to a strategic “human-in-the-flow,” ensuring that institutional accountability remains resilient in high-frequency, autonomous environments.
This article presents a deep learning-based mobile system for real-time detection of Diatraea spp. in maize crops. This pest, commonly known as the stem borer, causes significant economic losses in agriculture. Traditional control methods rely on manual identification and widespread pesticide use, often proving inefficient. To address this, a convolutional neural network-based on the YOLOv8 architecture was trained on a labeled dataset of field images from Ecuadorian maize plantations. The model detects the pest at various stages (egg, larva, adult) and identifies crop damage. Integrated into a mobile application via an application programming interface, it enables farmers to capture or upload images and receive immediate detection results. Experiments show over 90% accuracy with response times under 4 s, supporting real-time field use. This solution advances sustainable agriculture by improving early pest detection, minimizing pesticide overuse, and supporting informed decision making. It exemplifies the potential of mobile artificial intelligence in precision pest management.
Reliable webhook delivery is critical in large-scale financial technology systems, where delayed or failed notifications can disrupt fraud detection, transaction processing, and customer experience. Existing policies apply uniform pacing to all failure types, wasting capacity on unrecoverable failures and delaying recovery of time-sensitive events. This article presents a response-code-driven retry decision framework for production-scale fintech webhook delivery, in which each failure category receives distinct retry behavior—suppression, short-window retry, or front-loaded retries—rather than uniform persistence. The framework was parameterized from approximately one million anonymized production webhook delivery events and evaluated through replay-based simulation. The proposed approach preserved event-level delivery success rate (99.35%) while reducing median recovery latency from 10.0 to 1.0 min, p99 recovery latency from 40.0 to 3.0 min, and total retry load by 68%. These results demonstrate, in a single-platform case study, that category-aware retry logic can reduce retry load without degrading delivery reliability.
The presence of generative artificial intelligence (GenAI) tools represents a critical challenge in computing education. It affects academic integrity, blurring authorship, and rendering traditional, product-focused assessments unreliable. We present a collaborative and process-centric approach to mitigate GenAI misuse within a high-stakes Computer Science Group Project module. Instead of the reactive, unsustainable arms race of AI detection, we rely on a pedagogical redesign centered on three aims: redesigning assessment to demand verifiable process evidence and higher-order cognitive skills; establishing clear boundaries for responsible GenAI use as a professional, collaborative tool; and enhancing student success and inclusivity by leveraging AI to improve formative feedback. Module integrity is secured through a comprehensive Group Project Portfolio assessment, structured around three key components designed to mitigate GenAI dependency. Furthermore, the article reports on an initiative to use GenAI proactively to provide more inclusive and effective feedback, thereby reducing students’ reliance on AI for content creation.
The article serves as an artificial intelligence (AI) reality check for the C-suite. Executives must acknowledge that AI is not a trend but a permanent shift in the global infrastructure, and failing to act is a failure of leadership. The core of the article contrasts sanitized “textbook advice” with the “dark arts” of real-world technology adoption. While standard playbooks suggest polite upskilling and “Centers of Excellence,” advice from the “trenches” advocates for a hostile takeover of one’s own technology roadmap. This includes “bribing” middle managers, hiring outside AI “mercenaries,” and setting aggressive “kill dates” for legacy systems. The conclusion delivers a final, chilling message to Boards of Directors: The competition has moved past the AI “novelty” phase and into the “assassination” phase. The article ends with a stark ultimatum for senior management: Either weaponize AI to become the disruptor, or prepare to be the target of a competitor who already has.
This article presents and assesses the integration of an artificial intelligence (AI) agent into an educational escape room conducted in four different higher education courses on databases. The AI agent, implemented as a web-based chatbot using the GPT (Generative Pre-trained Transformer)-4o-mini model, was integrated through an open source platform to provide escape room participants with adaptive, process-oriented hints and support. A total of 148 students participated in this study. Results show that students positively valued the AI agent’s usefulness and ease of use, reporting that it helped them progress and reduce frustration. Performance data indicated fewer predefined hint requests and more efficient problem-solving compared to a previous edition of the same escape room without the agent. This study highlights the potential of AI agents as pedagogical assistants that foster engagement and autonomy in game-based learning activities.
This article argues that artificial intelligence (AI) is not mainly eliminating jobs but reducing the need for people to coordinate work inside firms. By taking over coordination tasks, AI reshapes organizational structures and reallocates labor toward oversight and higher-level functions.
Artificial intelligence (AI) has recently gained a mainstream presence. This presence suggests profound societal change. This article explores the technology underlying AI using the metaphor of a four-year course of study and beyond. It moves from technical to increasingly more speculative topics to illuminate the ultimate metaphysical AI question: whether AI will ever become conscious, much less omnipotent.
GenAI has reignited concerns about incipient unemployment. A more general question relates to creating meaningful lives if largescale unemployment materializes. The concept of a Civic Community (CC) is defined and suggested as a possible stabilizer for society in such a future. Characteristics of CCs are considered as well as conditions could support their emergence.
The increasing deployment of the Industrial Internet of Things (IIoT) in critical infrastructure sectors like manufacturing, healthcare, and transportation has shown new challenges for Digital Forensics (DF). Traditional DF methodologies are not well equipped to handle the complexity, scale, and heterogeneity of IIoT environments. This paper introduces a comprehensive Twelve-Step Process (TSP) tailored specifically for IIoT incidents, addressing the need for effective investigation and Potential Digital Evidence (PDE) handling in such dynamic environments in DF. We begin by exploring the importance of IIoT and its role in industrial ecosystems, followed by an examination of existing DF challenges. Each step of the process, from forensic readiness to investigation closure, is designed to ensure robust PDE collection, analysis, and legal compliance to increase chances of admissibility from a DF scenario
In contemporary IT organizations, the unauthorized and informal use of generative AI tools, commonly referred to as shadow AI, has emerged as a growing organizational risk. Although often treated as a technical or compliance issue, shadow AI is fundamentally shaped by human behavior and organizational culture. This article argues that low psychological safety is a key driver of covert AI use. When employees fear judgment, negative evaluation, or admitting uncertainty, they are less likely to discuss how they use AI tools in practice. Under these conditions, generative AI becomes a private coping mechanism that may increase hidden risks related to security, privacy, reliability, and accountability. Drawing on organizational and behavioral perspectives, this article examines how psychological safety influences AI-related behavior, identifies warning signs associated with shadow AI, and proposes a human-centered framework for responsible AI governance. We argue that transparent and effective AI management depends not only on technical controls, but also on creating environments where people feel safe to speak openly.
Existing frameworks, such as ISO/IEC 27001, offer limited support for the operationalization and validation of cybersecurity in critical IT/operational technology (OT) infrastructure, lacking procedural specificity and necessary attention to human factors in industrial environments. We present POSEIDON, a modular risk analysis framework that enhances current standards by embedding cybersecurity requirements throughout the entire technology readiness level (TRL) cycle. This innovation ensures security from concept right up to deployment. Furthermore, the framework expands the security lifecycle with TRL 10 for the secure decommissioning of long-life systems, which is crucial for the governance of critical assets. POSEIDON also incorporates cyberpsychology and cultural metrics to improve human resilience. Empirical validation within a complex naval IT/OT environment confirms its efficacy, demonstrated by the significant reduction in Mean time to detect, contain, and respond. POSEIDON offers a rigorous and transferable model for risk management and enterprise security.
Generative artificial intelligence (GenAI) is reshaping the software development landscape, introducing profound challenges to the professional identities of software developers. This study focuses on the identity threats posed by GenAI tools to software professionals and their effects on learning, skill development, and professional behaviors. Through an exploratory qualitative study, we investigate how software professionals perceive and respond to these challenges. Our findings reveal the nuanced ways in which identity threats shape their appraisals, emotions, and coping behaviors, potentially influencing professional growth and adaptation. Based on these insights, we propose four key recommendations: tailoring learning strategies, co-creating professionals’ identities, adjusting performance evaluation criteria, and fostering leadership advocacy. These recommendations aim to promote sustained professional development amid GenAI-driven transformations.
