Preliminary assessment of AI as a triage tool for forensic toxicology case interpretation data

Large language models (LLMs) such as ChatGPT have demonstrated potential for interpretation in various scientific disciplines; however, their application in forensic toxicology remains unexamined. We wanted to examine the performance of LLMs compared to experts at interpreting drug concentrations as a triage tool. In this study, a range of anonymised forensic toxicology case results from published sources were submitted as prompts to Microsoft 365 Copilot and ChatGPT version 3.5. AI-Generated outputs were assessed against the published expert interpretations for accuracy in drug identification, risk categorisation (fatal, life-threatening, severe, etc.), caveats to the interpretation (e.g. post-mortem redistribution), and expression of confidence (suggests, strongly suggests, etc.). Differences in the number of caveats given were observed for experts ranging from 0–5 per case (mean = 2.2) compared to Copilot (0–5, mean = 1.3) and ChatGPT (1–7, mean = 3.9). While results indicate that LLMs may assist in early triage under supervision, their use in evidential contexts is not currently supportable.