, hard-pressed pathologists presented with a corpse are able to take a bite of their sandwich and instantly pronounce a time of death. Reality is, of course, a lot messier, and the results—or lack of them—can make or break a case.
Now artificial intelligence is offering a helping hand. By analysing thousands of deaths and what follows, the technology can offer the best estimates so far of the post-mortem interval part of a forensic investigator’s work. For decades these specialists have had to rely on intuition, combined with observations of the state of the deceased and clues such as temperature, both of the dead body and the environment.
Forensic-science journals are full of such cases—some routine, others bizarre—while the potentially useful details of thousands more investigations are buried in case files around the world. Now forensic researchers in America are working to access and compile these valuable write-ups, and to use machine learning to analyse them.
To use the tool, forensic investigators attempting to date a body simply enter some specifics of their case into an app. Entering the location allows themodel to factor in local weather conditions. The investigators then add observations about the deceased, such as whether they were obese, whether insects have moved in and whether dogs or rats have gnawed on the body.
The results could be used to check alibis and help solve crimes, but they have other uses too. Madeline Atwell, a forensic anthropologist at Clemson University who works on the project , says the model has already helped close several missing-person cases. Combining time of death with when people were last seen alive is very useful in identifying human remains. “You match it with missing-person records, and that helps narrow your time frame,” she says.