Downloads: 1
Research Paper | Computer Science | Volume 15 Issue 4, April 2026 | Pages: 1165 - 1172 | India
Rethinking Workplace Safety: An Integrated AI System for Detecting Harassment Across Digital Communication Channels
Abstract: Rethinking Workplace Safety:An Integrated AI System for Detecting Harassment Across Digital Communication Channels is an AI-enabled multimodal workplace harassment detection system developed to identify abusive, discriminatory, intimidating, and unethical behavior across digital communication platforms such as emails, text messages, voice conversations, and video meetings. The system uses Natural Language Processing to analyse textual content and detect bullying, threats, manipulation, and inappropriate language. Speech processing techniques are applied to evaluate tone, pitch, intensity, and speaking rate for identifying verbal aggression. In addition, computer vision methods are used to study facial expressions, gestures, and visual cues in video interactions to detect non-verbal harassment. The system also considers conversation context, escalation trends, stress indicators, and organizational hierarchy to improve detection quality. Outputs from all modules are combined into a Threat Severity Score for early risk identification. The proposed system also supports multilingual analysis and interactive dashboards to help organizations maintain a safer digital workplace.
Keywords: Workplace Safety, Harassment Detection, NLP, Computer Vision, Speech Analysis, Multimodal AI
How to Cite?: Richu Mariyam Alex, Sindhu Daniel, "Rethinking Workplace Safety: An Integrated AI System for Detecting Harassment Across Digital Communication Channels", Volume 15 Issue 4, April 2026, International Journal of Science and Research (IJSR), Pages: 1165-1172, https://www.ijsr.net/getabstract.php?paperid=MR26419200919, DOI: https://dx.dx.doi.org/10.21275/MR26419200919