Abstract

Artificial intelligence-enabled safety technologies have proliferated across transportation, industrial, healthcare, and public infrastructure domains over the past decade. This systematic review and meta-analysis quantifies the measurable impact of AI-driven safety interventions on accident prevention and public risk mitigation. Drawing from 127 peer-reviewed studies, industry reports, we analyze effectiveness metrics including accident reduction rates, false positive/negative ratios, response time improvements, and cost-benefit outcomes. Our findings demonstrate that AI-enabled safety systems achieve an average 34.7% reduction in preventable accidents across analyzed domains, with significant variation by sector (range: 18-61%). Advanced driver assistance systems show the highest impact (42-61% accident reduction), while industrial predictive maintenance systems demonstrate 31-38% reductions in critical failures. However, implementation challenges including algorithmic bias, transparency deficits, and human-AI interaction failures partially offset these gains. We propose a standardized framework for evaluating AI safety technology effectiveness and identify critical research gaps in long-term reliability assessment and ethical deployment considerations.