• Researchers trained an AI model on blood metabolites from thousands of deaths. • The model predicts time of death with higher precision than traditional methods. • Development involved collaboration between universities and forensic agencies. • Findings could aid criminal investigations and medical examinations. • The study was published in Nature Communications.

Article Summaries:

  • Researchers at Linköping University and the Swedish National Board of Forensic Medicine have developed an artificial‑intelligence model that can estimate the time of death more accurately than current forensic methods. The AI was trained on metabolite profiles from thousands of blood samples taken from real deaths, allowing it to detect the biological changes that occur after death. The study, led by post‑doctoral fellow Rasmus Magnusson and published in Nature Communications, suggests the tool could aid murder investigations by narrowing the post‑mortem interval. The team notes that “death is a strong biological signal,” which the algorithm exploits to improve precision.

Sources: