Role of Artificial Intelligence (AI) in Medico-legal Practicing Aspect of Forensic Medicine

Abstract

Background: Artificial Intelligence (AI) has significantly impacted in Medico-legal Practicing Aspect of forensic medicine by enhancing diagnostic accuracy and improving investigative processes. However, its potential for transforming forensic analysis remains underexplored. Objective: To prospectively evaluate the role of AI in improving diagnostic efficiency and accuracy in Medico-legal Practicing Aspect of forensic medicine, focusing on medical image analysis for determination of age and detection of foreign body like bullet, pellet, gunshot residue etc. in deceased victim during postmortem examination, and toxicological interpretation. Methods: A prospective study was conducted on 64 deceased victims from the Department of Forensic Medicine & Toxicology, Rajshahi Medical College (RMC) as an autopsy sample, during the period of June 2023 to December 2023. AI tools, including machine learning models for image recognition, were integrated into routine forensic diagnostic procedures. Various forensic parameters, including post-mortem analysis, toxicology reports from CID forensic lab, and skeletal injury analysis, were processed using AI. Standard statistical methods, including t-tests, ANOVA, and chi-square tests, were used to analyze the data. Variables assessed included diagnostic accuracy, processing time, identification of toxic substances, and error rates in AI predictions. Results: The integration of AI tools improved diagnostic accuracy by 27%, with AI models identifying 95% of anomalies in medical imaging (standard deviation: ±3.5%) and 92% of DNA matches (p-value = 0.03). Toxicological analyses showed a 24% improvement in the identification of substances, reducing false negatives by 15% (p-value = 0.02). The mean processing time for forensic data reduced by 35% (p-value = 0.01), with a standard deviation in diagnostic time reduced from 3.2 hours to 2.1 hours (p-value = 0.04). The error rate in skeletal injury analysis dropped by 18%, with AI correctly identifying 88% of cases (p-value = 0.05). Furthermore, AI’s contribution in streamlining post-mortem reports reduced the time for documentation by 40% (p-value = 0.02). Sensitivity and specificity for toxicological analysis showed significant improvements, reaching 92% (p-value = 0.01) and 90% (p-value = 0.03), respectively. Conclusion: AI significantly improves both the diagnostic accuracy and efficiency of forensic medicine, offering valuable contributions in various forensic applications.

Ähnliche Arbeiten