Sherlock Holmes once famously quoted, “You see, but you do not observe.” To the growing list of people who merely see comes a boon in the form of AI, especially for the people who work in forensics. Using AI to detect blowfly activity, an AI tool can determine the time of death and uncover other details regarding the unnatural incident, even years later.
Blowflies are saprophytic, i.e., they feed on decaying organic matter. They lay eggs on dead bodies, and the larvae feed on them before growing into adults. These insects leave a signature remnant around the body—puparial casings. These reddish-brown compartments are where the larvae live during their growth phase. Analyzing these casings can help investigators back track to the time of death. They do so by estimating the time needed for the larvae to reach the casing stage.
Different species take different amounts of time to mature. To correctly determine the time of death, it is crucial to identify the species of blowflies the forensic team is dealing with.
Now, Rabi Musah, an organic chemist at Louisiana State University in Baton Rouge, and her team have combined an AI tool with chemical detection technology, allowing faster—just 90 seconds—identification of the blowfly species in question from its casings alone.
Time of death isn’t the only thing these casings could help determine, says Falko Drijfhout, an analytical chemist at Keele University in England who was not involved with the work. Casings can reveal whether the body has been moved. Casings generally remain close to the body, but if these are found far from the corpse, that indicates that the body has been shifted.
Blowflies are extremely proficient in detecting dead bodies. They can smell them from two kilometers away, just hours later. They lay eggs on the corpse. After hatching, the maggots feed on the corpse with their hook-like proboscis and eventually enter the casing stage. At this stage, they become the pupae, housed in the casings. Musah calls these sturdy, hardy little structures.
Casings of different species of blowfly look similar, and it is impossible to detect differences with the naked eye. One way to identify the species is to rear the offspring to adulthood and identify them once they have fully grown. This looks good on paper; however, oftentimes, investigators are left with only the casings. The other option involves DNA analysis; however, it can be fallacious, as DNA is prone to degradation under environmental stress.
The solution Musah developed has less to do with biology and is based mainly on chemistry. The team mapped chemical fingerprints of these casings, which are unique to each species. They used a technique known as field desorption-mass spectroscopy. The researchers then deployed a machine learning prediction model, a type of AI, to analyze the chemical data and identify the species of blowfly involved.
Musah trained her model on chemical footprints from hundreds of casings collected from lab-grown blowflies. The team then put their tool to a test, in which it had to identify 19 different species of blowflies from their casings, collected across the US. The AI never saw these casings before, but by finding out their chemical footprints, it was able to identify the blowfly species.
These chemicals might stick around for years, long after the person’s death, which the AI tool can leverage for chemical fingerprinting. The AI can also provide an idea of the time of death by noting the amount of weathering the chemicals have undergone.
These clues are like language and have a lot to say, while the AI tool is doing the linguistics. We are yet to explore further developments in AI use in forensics. Someday, AI can analyze blood splatter and determine the time of injury. These tools can also be tested on Frances Glessner Lee’s crime scene diorama—Nutshell Studies of Unexpected Deaths—which are still used in forensic training, in the future.
Contributors
Copyright @smorescience. All rights reserved. Do not copy, cite, publish, or distribute this content without permission.
