NYC shelter adopts AI tool to screen cats for pain

Animal Care Centers of NYC has begun using Sylvester.ai’s feline pain-detection platform during cat intake and medical assessments, marking a notable real-world deployment of AI in shelter medicine. The partnership gives ACC access to a smartphone-based tool that analyzes facial indicators linked to pain, with the company positioning it as a clinical support tool for fast-moving shelter workflows rather than a substitute for veterinary assessment. (myvetcandy.com)

The backdrop is a longstanding problem in feline medicine: cats often mask pain, and clinicians are still working against under-recognition in both general practice and shelter settings. The 2019 Scientific Reports paper that established and validated the Feline Grimace Scale helped formalize facial-expression-based pain assessment in cats, and AAHA pain guidance has since reinforced the need for standardized pain scoring in routine evaluations. In shelters, those challenges are amplified by stress, abbreviated handling, and the need to triage many animals quickly. (nature.com)

According to Vet Candy Radio’s report, ACC is using the system to flag possible pain in real time so teams can prioritize diagnostics, analgesia, and monitoring earlier in a cat’s stay. Sylvester.ai says its platform is built on facial analysis and machine learning and can be used in-clinic, through partner platforms, or by pet parents through a companion app. The company has also been building its shelter-data footprint elsewhere: in November 2024, it announced a data-licensing agreement with Austin Pets Alive! to use anonymized shelter-cat data to refine its models. A separate 2024 grant announcement from CATalyst Council said Sylvester.ai had processed more than 350,000 images, suggesting the company is trying to scale beyond a single-use pilot. (myvetcandy.com)

There’s also a broader research arc behind this. Beyond the original grimace-scale work, later research has explored fully automated deep-learning models for smartphone-applicable feline pain prediction, showing the field is moving from manual scoring toward image-based automation. That doesn’t mean the science is settled. Even supportive sources frame these tools as adjuncts, and the literature itself notes the limitations of existing pain scales and the confounding effects of demeanor, drugs, and context on feline pain scoring. (nature.com)

Direct outside commentary on the ACC rollout appears limited so far, but the industry framing is consistent. Fear Free’s preferred-product materials have highlighted Sylvester.ai’s alignment with veterinary pain-assessment protocols and repeated the company’s 89% accuracy claim, while AAHA educational materials continue to emphasize that feline pain assessment should combine hands-on evaluation with structured tools such as the Feline Grimace Scale. Taken together, that suggests cautious interest: the profession is open to better standardization, but still expects pain tools to support, not replace, veterinary interpretation. (fearfreepets.com)

Why it matters: For shelter veterinarians and technicians, the real significance is operational. A photo-based tool could help create more consistent intake screening across shifts, experience levels, and crowded caseloads, particularly for cats whose pain signals are easy to miss when fear and stress muddy the picture. If the system also follows the cat into the adoptive home, it may extend the clinical conversation beyond discharge and give pet parents a structured way to notice changes that warrant recheck. That could be especially useful for recently sterilized cats, animals with dental disease, orthopedic pain, or unresolved medical concerns leaving the shelter. This is partly an inference from how the tool is described, rather than published outcomes from ACC, but it fits the workflow problems shelter teams routinely face. (myvetcandy.com)

There are still open questions. ACC’s own recent reports show an organization expanding facilities and capacity, but public data on this AI deployment’s clinical impact haven’t surfaced yet. Veterinary professionals will want more than a precision figure: they’ll want to know sensitivity, false-positive rates in stressed shelter cats, how results compare with clinician scoring, and whether the tool changes treatment decisions or outcomes. They’ll also want clarity on training, image quality requirements, and how the platform performs across coat colors, face shapes, ages, and brachycephalic cats. (nycacc.org)

What to watch: The next meaningful milestone will be prospective shelter or clinic data showing whether AI-assisted screening improves pain recognition, speeds intervention, or supports better continuity of care after adoption, and whether ACC or Sylvester.ai publishes those findings in a peer-reviewed or conference setting. (myvetcandy.com)