UF explores AI to identify suspicious cat deaths

The University of Florida is using veterinary forensics and artificial intelligence in an effort to better identify suspicious cat deaths. In February 2026, UF highlighted work by Adam Stern, DVM, and Jon Kim on an AI tool meant to detect patterns that help determine whether a human was involved in a cat’s death, extending the mission of UF’s Veterinary Forensic Sciences Laboratory to investigate cruelty in free-roaming cats. (vetmed.ufl.edu)

The effort builds on several years of work inside Stern’s “A Cat Has No Name” program, which was created to investigate deaths in free-roaming cats and document findings that could support animal cruelty cases. UF’s forensic lab says many deceased community cats would otherwise be collected and disposed of without examination, leaving veterinarians, shelters, and authorities with little insight into underlying disease, trauma patterns, or possible abuse. The program has also evolved beyond necropsy alone, including disease surveillance and field response when clusters of cat deaths raise concern. (animalforensics.vetmed.ufl.edu)

That history matters because Stern has repeatedly framed forensic investigation as a corrective to assumptions. In a 2020 UF profile, he noted that animals found dead by roadsides are often presumed to have been hit by vehicles, even though only a full investigation can establish cause of death. In prior cases, the lab identified projectile injuries in dogs and cats and referred suspected cruelty cases to law enforcement; UF has also said some of those cases moved into the judicial system. Stern’s work has since expanded into broader cruelty-response infrastructure, including coordination of animal cruelty task forces in Florida and courtroom testimony in multiple jurisdictions. (vetmed.ufl.edu)

The new AI angle appears aimed at one of the hardest parts of veterinary forensic work: turning many small observations into a reliable signal. UF’s lab describes its cat and dog death investigation programs as efforts to understand mortality events, educate trainees, and provide law enforcement with usable evidence when cruelty is suspected. Coverage of the project says Stern has spent years building a repository of case data and refining how those data are collected, with the AI tool intended to identify repeatable patterns associated with human involvement. Based on UF’s description and outside coverage, the likely value is not replacing the pathologist’s judgment, but helping standardize case review and flag combinations of findings worth closer scrutiny. That last point is an inference from the stated goals of pattern detection and standardization. (animalforensics.vetmed.ufl.edu)

Industry reaction so far is limited, but UF has consistently positioned Stern as a leading voice in veterinary forensic science. In 2025, the Animal Law Section of The Florida Bar recognized him for contributions to animal welfare in Florida, citing his research on mortality patterns in free-roaming cats and his role in promoting evidence-based animal protection. Outside trade coverage has echoed that framing, describing the cat-death project as an attempt to use AI to support a growing forensic dataset rather than as a standalone diagnostic system. (cdpm.vetmed.ufl.edu)

Why it matters: For veterinary professionals, especially those in shelter medicine, pathology, community cat programs, and animal welfare, this work sits at the intersection of public health, cruelty detection, and clinical documentation. Better forensic pattern recognition could help distinguish infectious disease outbreaks from intentional harm, support more defensible reporting to authorities, and improve how cases are recorded across institutions. UF has already shown the practical overlap: its cat forensics program helped investigate a South Florida colony with a nearly 50% mortality rate and ultimately identified bacterial bronchopneumonia and severe hookworm-associated anemia rather than poisoning. That kind of differentiation can shape treatment decisions, colony management, and whether a case becomes a public health response, a welfare intervention, or a criminal investigation. (veterinarypage.vetmed.ufl.edu)

There’s also a training implication. UF says its forensic programs are used to teach veterinary students and residents how to perform forensic death investigations, and the lab’s broader research spans toxicology, postmortem chemistry, and minimally invasive autopsy methods. If AI-supported pattern analysis matures, it could eventually become part of how future veterinarians learn to document suspicious injuries, interpret postmortem findings, and communicate with animal control or prosecutors. (animalforensics.vetmed.ufl.edu)

What to watch: The next key milestones are whether UF or its collaborators publish peer-reviewed validation data, clarify the size and makeup of the dataset behind the model, and show how the tool performs in real-world case triage. Veterinary professionals will also want to know whether the system is being built primarily for internal decision support, multi-site forensic collaboration, or eventual use by shelters, animal control, and law enforcement partners. (vetmed.ufl.edu)