As VIN News’ fantastic reporting covered, a veterinarian in Bartow, GA was surprised to see her own clinic featured in a viral TikTok video showing staff caring for puppies. Although initially startled, she was happy (and not surprised) to see that the recording portrayed her team acting professionally and compassionately. While the online sentiment online for how the vets acted was positive, the incident underscores a broader trend: clients increasingly record veterinary visits using smartphones, smartwatches, and now smart glasses. Recording consultations, whether openly or covertly, is becoming more common across healthcare settings, with human medical studies showing meaningful percentages of patients recording visits, sometimes without permission.

Here’s the video:

Clinics must assume they may be recorded at any time, with footage potentially edited or shared widely online. While recordings can validate professionalism and enhance transparency, they also pose risks to staff morale, client privacy, and contextual accuracy. Some practitioners see value in permitting or even initiating recordings for client education, documentation, informed consent, and training, including the use of AI-powered scribing tools. In human medicine, some products are coming to market as a scribe for the patient. It is only a matter of time until these make their way to veterinary medicine, which will reinforce considerations around who records what in the exam room.

Other practices are considering clearer boundaries, signage, or intake form disclosures to manage expectations as wearable recording technology becomes less visible and more prevalent.

Legally, the permissibility of secret recording depends on jurisdiction and whether the interaction occurs on private premises. In the United States, most states follow one-party consent laws, meaning only one person in a conversation needs to consent to and be aware of a recording (presumably the one doing the recording). Still, practices do have reasonable authority to set policies governing recording within their clinics. As recording becomes more common, if you’re in a position to do so, consider developing written policies on audio and video recording that is proactively shared with staff and clients. First, evaluate your local state or provincial law, be sure to address employee privacy considerations, and ensure compliance with confidentiality obligations when other clients may be captured.

New findings from the American Veterinary Medical Association 2025 Pet Ownership and Demographic Sourcebook show that most U.S. pet owners report high satisfaction with their regular veterinary practice.

More than 60 percent of clients say they are extremely satisfied with staff friendliness, staff knowledge, and overall quality of services. Even in traditionally sensitive areas such as cost, roughly two-thirds of owners report being extremely or somewhat satisfied. For veterinarians, this reinforces that communication, compassion, and service quality are being recognized and valued more than teams may assume.

At the same time, the data identifies practical growth areas. Satisfaction scores were lower for cost of services and hours of operation, both of which influence visit frequency, treatment adherence, and long-term loyalty. Based on these findings, the AVMA has some recommended actions for clinics and vets:

Last week I wrote about how Claude Co-work represents a massive shift in AI. After another week of exploring this AI agent, it is only more clear that things are becoming dramatically different.

Anthropic likely released Co-work in part because of the incredible popularity of OpenClaw - an open source AI personal assistant built by Austrian developer Peter Steinberger. While some have used OpenClaw to run their life or business, it was full of security flaws and opened the user up to vulnerabilities. It wasn’t suitable for the average person, and even dangerous for some of the sharpest developers. Claude’s release represents a professionalization of that same idea. More guardrails, higher levels of security.

But Claude is not the only one moving in this direction. OpenAI hired OpenClaw’s creator in mid-February and this week Microsoft announced CoPilot Tasks. This means that for all the major companies, AI won’t just speak, it will act on your behalf.

This is not entirely new. With enough time and energy connecting LLMs to external tools like Zapier, Make, and n8n, many things have been able to be automated with AI in the past few years. However, these new AI systems democratize this ability to non-technical users and expand their capabilities.

The capabilities of these systems are likely to grow. Right now, the internet has been built to limit the abilities of computers to do things on their own. Think of it this way: CAPTCHA and reCAPTCHA exist to force users to prove they are human, but these guardrails are changing.

At the end of last week Google and Microsoft released WebMCP, a protocol which allows developers to open their websites to AI agents. This means that soon more of the web will be easier to navigate for AI agents, increasing their abilities and removing the requirement of human oversight for certain actions. Given that some people believe that artificial general intelligence is here, giving agents these abilities/tools is a dramatic shift in how the world will work.

Much of this might seem scary to you. I often oscillate between excitement and absolute dread, sometimes in the same conversation. But I found the analogy offered by Boris Cherny, the creator of Claude Code, insightful and hopeful. When speaking on Lenny’s Podcast last week he likened this shift in coding ability to the advent of the printing press. Before the printing press, literacy was low and reserved for the elite. But afterwards, humanity thrived and printed texts and knowledge translation was a huge part of that. The printing press democratized this ability to communicate with the written word.

In the same way, AI-assisted coding can democratize software development. The ability for AI agents, powered by these coding models, to take over tasks is about to alter the way we work.

In 2022 I adapted the words of radiologist Curt Langlotz and wrote in JAVMA that “Veterinarians who use AI will replace veterinarians who don’t.” But in all honesty, this was not exactly what I had in mind. I thought that computer-aided diagnostic systems would develop in such a way that veterinarians would need to use them to remain relevant in the field. I still believe this, however the impact of those systems is not yet as profound as I believe this shift to agentic AI is about to become.

-RBA

Here are some of the other stories that caught our eye and we're following this week from around the veterinary world and animal kingdom: