From Guidelines to Practice: What 19 Veterinary Leaders Are Saying  About Responsible AI

Chances are, AI is already part of your veterinary practice, whether it’s being used with social media tools, in the companion animal exam room, or in a practice barn call. Generative AI scribes are transcribing appointments in small animal clinics nationwide, and pet owners increasingly show up having already consulted AI about their animal’s symptoms. The technology is moving faster than any other technology in veterinary medicine to date, while the profession’s shared norms for using it responsibly are still catching up, with direct implications for how veterinarians document care, disclose AI use, and manage client trust.

That gap is the focus of a new IRB-approved qualitative study I conducted with 19 leaders in veterinary medicine, building on a Phase 1 poster from the Symposium on Artificial Intelligence in Veterinary Medicine (SAVY) in 2025. The expanded research draws on 19 in-depth interviews plus nine background conversations with veterinarians, PhDs, C-suite executives, AI patent holders, regulatory staff, and association leaders across practice, academia, industry, and government. Participants averaged 21 years of experience.

The result is a communications perspective on how veterinary medicine, amid variances in state laws and board guidance and with no unified federal law governing AI use, is working out, in real time, what “responsible AI” actually looks like in daily practice.

No Rulebook, But Not a Free-for-All

One veterinary leader captured the current moment bluntly: “We are in the phase [of AI]where you can still drive and text at the same time.” The point isn’t that AI use is unregulated; it’s that veterinarians may not realize existing laws and standard-of-care practices already apply. Ignorance of that fact, the source warned, may not hold up as a defense in a future disciplinary case or lawsuit.

The research frames this as a patchwork rather than a vacuum. There’s no single federal framework, and board guidance varies widely, but reference points exist: the Veterinary Model Practice Act offers a baseline, and the veterinary oath anchors ethical obligations above any tool. Responsible AI isn’t waiting on new legislation; it’s an extension of obligations practitioners already carry.

Finding 1: Scribes Are Driving Adoption, But Trust Barriers Remain

The study’s first finding is that AI adoption is fragmented, but generative AI scribes have become the clear accelerant. Diagnostic AI has been used for years in radiology, but transcription and documentation software, tools that sit quietly in the exam room, are pulling the broadest swath of the profession into AI use today.

The appeal is straightforward: reduced after-hours paperwork and improved well-being. As one mid-career veterinarian and clinic AI implementation manager put it, paperwork routinely ranks among the top three reasons veterinarians burn out.

Adoption isn’t uniform, though. Barriers remain real: skepticism, uneven AI literacy, absence of shared standards, and the difficulty of getting full team buy-in. A boarded associate director at a small animal teaching hospital said the intent isn’t to cut corners: “We’re using them to practice better medicine and help more pets.”

Finding 2: Disclosure Is Contested, And A Risk-Tiered Model Is Emerging

Perhaps the most consequential finding concerns disclosure. There is no universal consensus on whether, or how, veterinarians should disclose AI use to clients, even though disclosure is broadly advised. Documentation of AI involvement is similarly unsettled, with real implications for credibility and client trust.

What is emerging is a risk-tiered approach: expectations for transparency intensify when recording, data capture, or clinical decision-making support is involved. The lower the stakes, the more disclosure norms diverge; the higher the stakes, the more consensus solidifies around the need for transparency.

One boarded veterinary radiologist and AI patent holder, with 43 years of experience, offered a caution: “We’re recording everything, what the client says, what the kids say. It’s recording everyday life.” The bigger risk, the source noted, is complacency: letting AI-generated transcriptions flow into medical records unproofread. The professional norm holds regardless of the tool: license, judgment, and final accountability remain human.

Finding 3: Pet Owners Are Already Using AI, Whether Vets Know It or Not

The third finding turns toward the exam room door. Pet owners increasingly use AI to self-diagnose, for themselves and their pets, often without telling their veterinarian, a trend that has intensified since 2025 alongside rising AI use in human healthcare.

One boarded practicing veterinarian and veterinary leader, with more than 25 years of experience, described the shift: “No longer do we have Dr. Google. No longer do we have Nurse TikTok.” Clients now walk in having already consulted ChatGPT, sometimes requesting a specific antibiotic. Rather than treat this as a threat, the veterinarian reframed it: “This is an empowered, educated owner. How can I take where they’re at and help them?”

That reframing captures a central thread: the profession’s opportunity isn’t to resist AI, but to stay positioned as the trusted expert who knows the specific patient.

What Comes Next

The research points toward concrete next steps: separating guidance for administrative AI use, like scribes, from diagnostic AI use, and situating practitioner perspectives alongside guidance already published by AAVSB and AVMA. At the practice level, that means standard operating procedures for reviewing AI outputs, recording, documentation, and vendor due diligence. On education, it means training veterinarians to fact-check AI tools, understand their limitations, and evaluate AI-influenced information clients bring in.

As the study author, I acknowledge real limitations: the sample skews toward leadership and change agents in the U.S. and Canada, and the AI landscape evolved even during the research period. Future research should bring pet owners’ perspectives into the picture and track how guidelines diffuse across day-to-day practice.

The throughline from 19 interviews with the profession’s AI-engaged leadership voices is consistent: the tools are already here, and use will keep expanding. The question isn’t whether to use them, but how, and members don’t have to wait for formal guidance. Reviewing AI-generated notes before they hit the record, having a clear practice policy on disclosure, and staying current on evolving AAVSB and AVMA guidance are practical starting points available today.

Lea-Ann Germinder, Ph.D., is president of Germinder & Associates, Inc., publisher of Goodnewsforpets.com, and recently received her doctoral degree in strategic communication from the University of Missouri School of Journalism. This article draws on research presented by Germinder at the Symposium on Artificial Intelligence in Veterinary Medicine (SAVY 3.0) at Cornell University College of Veterinary Medicine, May 30, 2026. The presentation “From Guidelines to Practice: Communicating Responsible AI in Companion Animal Veterinary Medicine” was awarded top oral presentation. Claude Sonnet assisted with the editing of this article, but the research and the work are the author’s own. A full journal article is expected to be published in late 2026.

Dr. Germinder can be reached at lgerminder@germinder.com or at 917-334-8682.

© Lea-Ann Germinder. All rights reserved.