Workplace accidents in high-risk environments remain a major safety concern, particularly when workers’ visual and auditory channels are overloaded. Haptic feedback offers a promising alternative for alerting individuals to unseen dangers and enhancing situational awareness. Motivated by challenges commonly observed in construction, this study investigates haptic alerting strategies applicable across dynamic, attentionally demanding contexts. We present two empirical experiments exploring how wearable vibration cues can inform users about approaching objects outside their field of view. The first experiment evaluated variations of pattern-based vibrations to simulate motion and examined the relationship between signal parameters and perceived urgency. A negative correlation between urgency and pulse duration emerged, identifying a key design factor. The second experiment conducted a novel comparison of pattern-based and location-based haptic alerts in a complex virtual environment, with tasks designed to simulate cognitive engagement with work processes. Results indicate that location-based alerts were more efficient for hazard detection. These findings offer insights into the design of effective user-centred haptic-based safety systems and provide a foundation for future development and deployment in real-world settings. This work contributes a generalisable step toward wearable alerting technologies for safety-critical occupations, including but not limited to construction.
8 December 2025 in Human-Intelligent Systems Integration
The application of Large Language Models (LLMs) in academic research faces unique challenges of privacy and workflow integration. This paper introduces TAUCHI-GPT, a novel, open-source AI assistant whose evolution informs our analysis. We detail its two versions: a cloud-based V1 using GPT-4 and reflection cycles, and a local, privacy-preserving V2 with RAG architecture. Based on empirical findings from two user studies, we present a critical Human-System Integration (HSI) analysis of the security vulnerabilities and alignment challenges inherent in local LLM deployments. We examine how recent development trends—such as model distillation and reward-model learning—and the complexities of internal model mechanisms exacerbate risks like prompt injection, RAG data failures, and unfaithful explanations that impact user trust. Drawing from HCI principles and mechanistic interpretability insights, we propose and discuss a multi-layered mitigation strategy. This work contributes significantly to HSI and AI by presenting an evaluated system, a rigorous analysis of local deployment risks from a sociotechnical perspective, and actionable, stakeholder-specific guidelines for the secure and responsible utilization of LLMs in academia.