Imagine a world where spays and neuters are performed by robots. Where pet owners with questions and concerns can engage with virtual employees 24 hours a day. Where medical records are generated automatically after each appointment. Where each veterinarian has a digital version of themselves handling mundane, time-consuming tasks while the human spends more time with patients and clients, sees more pets, and gets home earlier to enjoy a healthier work-life balance.
It might sound like science fiction, but all this and more could be possible someday for a veterinary profession incorporating artificial intelligence.
Dr. Cody Creelman, CEO of the Alberta, Canada, veterinary practice Fen Vet, has prepared for the paradigm shift. Launched in 2021, Fen Vet was built on a foundation of innovation — something Dr. Creelman said has been part of any industry that’s been reimagined.
To illustrate his point, he told a story about the business of ice:
“For many years, men cut ice off lakes and shipped it around the world to be put into iceboxes or used in manufacturing or factories. Then, someone discovered that ice could be made in a factory by pumping ammonia through pipes and pouring water over the top. Suddenly, you had a whole bunch of men standing on a frozen lake saying, ‘There’s no way this is going to take off because my product is the best.’ And what happened to those men? Their business disappeared. Then someone discovered you can make ice at home, and all the people in the factories making ice were saying, ‘There’s no way people are going to make ice at home. That’s crazy.’ But the business of making ice in factories was done. Those paradigm shifts were caused by innovation.”
After graduating from the University of Saskatchewan veterinary school in 2011, Dr. Creelman would go to sleep every night full of anxiety and a singular thought: “How am I going to be on the other side of that shift? How am I going to ensure I’m not the guy standing on the lake?”
Dr. Creelman believes that if you’re actively looking for that paradigm shift, you’re less likely to be left standing on the lake. He stays at the forefront of emerging tools and technology to constantly find the best way to do things. “And there’s some really great AI-enabled technology that allows us to execute on that,” he said.
What Is AI?
According to the Merriam-Webster dictionary, artificial intelligence is the capability of computer systems or algorithms to imitate intelligent human behavior. Generative AI is capable of generating new content, such as images or text, in response to a submitted prompt, or query, by learning from a large reference database of examples.
OpenAI’s ChatGPT is an example of generative AI. You can ask it to write a client email and then have it edit the response so the text sounds friendlier and more approachable. You can tell ChatGPT it was wrong about something, and it will respond with something along the lines of, “You’re right. I’m sorry I missed that.”
Engaging with ChatGPT isn’t like engaging with a computer. It feels like communicating with a human — an extremely smart, fast-typing human.
Reports have emerged of high school and college students submitting papers written entirely by artificial intelligence. One pet owner’s Twitter thread went viral when he claimed AI saved his dog’s life after a veterinarian couldn’t diagnose her correctly.
“This stuff is on the same level of innovation as search engines, broadband internet and smartphones,” said Dr. William Tancredi, a veterinarian and the owner of Old Ridge Veterinary Hospital in Chadds Ford, Pennsylvania. “If we ignore it, we’re going to be like the folks who refuse to send emails or still have a fax machine in their office. It’s the way the ground beneath our feet is shifting, and we don’t have a lot of say about it. We can either get the hang of it and use it or get knocked to the ground as it moves.”
While AI’s capabilities are impressive, there are concerns about it being used nefariously and its potential impact on human jobs.
“If you start diving deeper into the engineering side of things, you’ll see that AI isn’t built to replace humans,” said Sebastian Gabor, the founder of Digitail, a cloud-based practice management software that incorporates AI to automate backend processes throughout a pet owner’s journey with a veterinary practice. “However, someone using AI will replace someone not using AI.”
Gabor, a tech professional who spent 12 years building platforms and apps for the likes of The Walt Disney Co. and Hasbro, launched Digitail in 2018 after an unfortunate experience with his puppy Neo’s veterinary clinic. Today, practices in 16 countries use Digitail software.
How Veterinary Practices Can Use AI
1. Client Communication and Engagement
“We did some research, and 80% of the inbound phone calls to the front desk are the same 20 questions,” said Dr. Steve Merchant, a New Zealand veterinarian and the founder of Virtual Vet Nurse. “So, we thought, ‘Why don’t we just script those 20 questions and create a digital assistant that’s there to purely work for the team and take some of that load off?’”
Virtual Vet Nurse is an AI-powered conversational chatbot, or digital assistant, that allows clients to engage with their veterinary practice 24/7 while freeing up staff time to engage in clinical responsibilities. Virtual Vet Nurses are at work in clinics in the United States, Australia, New Zealand, South Africa, the United Kingdom, Southeast Asia and Canada. Each practice’s digital assistant is customized with a name, gender and persona. The virtual assistant automates and streamlines front-desk tasks, including scheduling appointments, ordering refills, obtaining pre-visit histories and information, and conducting post-op follow-ups. The scripts are specific to the veterinary practice, so Virtual Vet Nurse knows how to answer client questions the way the practice prefers.
Dr. Creelman’s Fen Vet team uses Digitail. He’s excited about what’s soon to come.
“In the very near future, someone will call our phone service, and the Fen fairy, Tails, will answer the phone,” he revealed. “The customer will have no idea — they’ll have a great conversation, their appointment will get booked, and their estimate will get sent over.”
ChatGPT also can be used for translations in person or via email or text. Dr. Creelman had a Ukrainian client whose dog was diagnosed with an ear infection.
“I asked ChatGPT to tell the pet owner that his dog had a bilateral ear infection caused by ‘this’ and that we’re going to treat it with antibiotics,” Dr. Creelman said. “I quickly reviewed ChatGPT’s response and then asked it to translate into Ukrainian, which it did instantly. Then I turned my computer screen to the client. He read it and smiled. He fully understood.”
Dr. Creelman’s team also uses artificial intelligence to generate post-visit appointment summaries and discharge notes for clients. One doctor on his team is “taking it to the extreme by extracting a bit of personalized information from the visit and getting AI to generate a custom-made client communication that incorporates the information in a whimsical sort of story,” he said. “It still includes all of the pertinent information, but it’s written almost like a message from your best friend who’s trying to make your day a little better. It adds a layer of creativity, and clients love it.”
In addition, AI can automatically survey clients after certain types of visits to help build your online reputation, conduct post-op follow-ups so you don’t play phone tag, and write email and social media copy to help market your practice.
THE POSSIBILITIES
- Prescreen and triage
- Schedule appointments and send reminders
- Gather patient histories
- Create discharge notes and educate clients
- Communicate in different languages
- Send invoices and estimates, and collect payments
- Solicit client feedback through surveys and online reviews
- Follow up after visits, procedures and hospitalizations
- Create marketing campaigns
2. Patient Care
Because AI can access veterinary textbooks, it might suggest tests to rule out or confirm specific diagnoses. And, while AI radiology and cytology interpretation aren’t new, the science is so accurate that “someday soon it will likely be considered malpractice if you don’t use it to confirm your diagnosis,” said Clint Latham, the founder of Lucca Veterinary Data Security.
But it doesn’t stop there. AI’s potential for helping veterinarians create the most successful treatment plans is something that excites even self-proclaimed big-tech contrarian Latham.
“Imagine you just diagnosed a Yorkie with transitional cell carcinoma and you ask AI to provide the treatment plans from the doctors who had the greatest success,” Latham said. “In an instant, it will scan billions of records all across the world. Now, you’ve been able to ‘interview’ a billion doctors and get the treatments that have been the most successful. And it happens instantly.”
Before going into an appointment, Dr. Creelman can ask Tails about his patient’s medical history. Tails provides a summary.
“Instead of scrolling through 14 pages to find out the patient had a rabies vaccine in 2019, I can ask my practice management software, and it tells me instantly,” he said.
AI’s ability to assist with medical records really excites Dr. Creelman. “We’ve been bogged down by cumbersome medical recordkeeping tasks that make workflow extremely challenging,” he said. “As an industry, we’re talking about a veterinarian shortage. I don’t believe that’s true; there’s an efficiency problem. If we could have all the vets that exist today just do medicine and surgery and not have to worry about the medical record component, we could literally do 200%, 300% or even 400% more. We could go faster, we could go harder, we could see way more patients, and we would have more fun doing it.”
THE POSSIBILITIES
- Diagnostic testing
- Treatment plans
- Medical records
3. Human Resources
At Fen Vet, Dr. Creelman uses ChatGPT to generate internal communications and all the paperwork needed for a more formal email. “It saves so much time,” he said.
THE POSSIBILITIES
- Job postings
- Job offers
- Letters of recommendation
- Job descriptions
- Employee handbook
The Benefits of Using AI
1. Save Time and Improve Efficiency
“When we talk about automation, we talk about the three D’s: dirty, dull and dangerous. We want to automate those things away,” said Chris McComb, an associate professor of mechanical engineering at Carnegie Mellon University. He leads the university’s Human+AI Design Initiative, a group focused on “figuring out how to bring together humans and AI agents in a way that optimizes our performance as well as our overall experience and joy,” he said.
In veterinary medicine, some of the most time-consuming tasks can also be the most dirty, dull and dangerous. Dr. McComb said AI can free up staff time to focus on a job’s more human portions.
“I don’t want AI euthanizing my pet,” Dr. McComb said, “but I do want AI taking care of pre-screening clients or scheduling appointments or keeping medical records so the veterinarian has more time to spend with me and my pet.”
According to Gabor, Digitail’s pre-screening, medical support decisions and client education have been the most powerful and timesaving for his customers.
Drs. Creelman and Tancredi are looking forward to AI’s growing impact on medical recordkeeping.
“A 2016 study revealed that 51% of physician time was spent on medical records instead of with patients,” Dr. Tancredi said. “If veterinarians’ numbers are similar, we can solve the so-called productivity crisis with AI.”
2. Enjoy a Healthier Work-Life Balance
At Old Ridge Veterinary Hospital, Dr. Tancredi’s team members save huge amounts of time thanks to AI. But instead of them seeing more patients, Dr. Tancredi wants to focus on helping his employees enjoy a better work-life balance.
“I don’t want AI augmentation to mean, ‘Oh, vets are seeing 20 patients a day, but now you need to see 40,’” he said. “Maybe it’s possible to see 20 patients a day in half as much time and we get parts of our lives back that we didn’t before.”
3. Improve Relationships and Experiences
When AI handles mundane, time-consuming, repetitive tasks, veterinary professionals can spend more time with clients and patients, improving pet health and customer loyalty.
“In the very near future, vets will be able to focus on their relationships with pet parents and patients, and rely on AI to find information quickly, compute a high volume of data and more,” Gabor said.
Dr. Tancredi likes that ChatGPT’s client email copy is kinder and more compassionate than what he might write as a “fairly direct human from the East Coast,” adding, “I like that ChatGPT remembers the social niceties at the beginning and the end.”
4. Practice Better Medicine
Dr. Tancredi uses AI daily in his practice and writes about it on his Substack account (substack.com/@drtanc). In an article titled “Centaurs, Cyborgs and How to Get Caught Up on AI in Vet Med,” he links to a Harvard Business School study and writes: “The immediately apparent advantage of generative AI is its potential to raise the quality of care provided by the lowest performers. And research has borne out that instinctive conclusion. Consultants in the study with below-average performance improved by a shocking 43% when using AI, with above-average performers improving by 17%. … If better medicine isn’t reason enough to explore something, you may be in the wrong line of work.”
Dr. Creelman and his Fen Vet team use AI-enabled radiology interpretation. “We have the ability to double-check and ensure we aren’t missing anything,” he said. “It’s enabled us to practice better medicine and to be faster, more confident veterinarians.”
Concerns About AI
The artificial intelligence landscape is evolving so fast that some tech leaders are calling for more oversight and regulation. Before you jump on the AI bandwagon, keep these considerations in mind.
1. Data Privacy
“If it’s free, you’re the product,” said Latham, of Lucca Veterinary Data Security. “No one really knows how ChatGPT has been building its dataset.”
Digitail has an enterprise agreement with OpenAI, according to Gabor. “When we ask a question, we need to send the data, but they’re obliged to delete it after the answer is finished, so they can’t store it, and they can’t then use it for training purposes. So, the data is brought back into the hands of the clinic,” Gabor said.
Here are tips to protect your data in an AI world:
- Read the privacy statement to determine what a company will and won’t do with your data. “If there are ambiguities or third-party sharing, those are probably red flags,” Latham said.
- Never use the notes section of your PIMS to store social security or credit card numbers.
- Don’t put anything into a large language model like ChatGPT that you wouldn’t give to a random high school student or post on social media.
2. Prompt Engineering
“When Google came out, we had to learn how to search online, and some people were better at it than others,” Gabor said. “With these AI tools, you need to know how to phrase the question or prompt to get the answer you’re hoping for.”
3. Source Material
Many AI sources are trained on almost the entire internet, which can make them less than accurate. “But you can fine-tune it by indexing the credible books and content that matter,” Gabor said. Dr. Merchant said his Virtual Vet Nurse introduced a curated GPT model that looks at only verified websites and data.
4. “Hallucinations”
Large language models can get confused and make up information, “and we don’t have any great procedures for negating those hallucinations entirely,” said Carnegie Mellon’s Dr. McComb. For example, an LLM could confuse the fact that chocolate is toxic to dogs and instead say, “Give your dog chocolate.” And the LLMs will almost always be confident in their accuracy.
“LLM systems can give really terrible recommendations,” Dr. McComb said. “But in the hands of experts who are fact-checking and know the right questions to ask, it’s safer.”
5. Reviews and Edits
“We scan and fact-check everything,” Fen Vet’s Dr. Creelman said about AI’s output. “We see errors often, but that’s expected. Editing does need to happen, but it truly and remarkably has increased our workflows.”
6. Liability
“If somebody uses a large language model or AI agent to design a bridge and the bridge falls down, who’s liable?” Dr. McComb asked. “At the end of the day, the responsibility still needs to fall with the human expert.”
7. Who’s in Charge?
AI has generated images and content considered racist, misogynistic, bigoted or politically biased. In reporting this article, I tested AI by entering two prompts into OpenAI’s free GPT-3.5. All politics aside, you be the judge on whether the large language model is politically biased.
The Future
What do the fast-and-furious AI advances mean for veterinary medicine in the years ahead?
“I saw a video of AI connected to a camera and one of those DaVinci-style robots with surgical capabilities dissect the shell off an egg, slit open the inner membrane and suture it back up,”
Dr. Creelman said. “Do you know how much harder that is to do than neutering a dog? There’s a company in China that’s doing AI blood draws. You put your arm in a device that scans where your vessels are, and then it shoves a needle into your vein and draws blood.”
Similar advances might never come to veterinary medicine, but we do know that AI is poised to change the profession in a number of ways over the next few years.
Digitail’s Gabor believes that AI will become embedded at every step of patient care within 12 to 36 months. He also thinks artificial intelligence will force everyone to better understand the importance of data.
“It’s hard to leverage a lot of that data in the right way,” he said. “So, there will probably be some force factor in improving data quality and data education and creating some alignment in the veterinary market.”
Looking further into the future, Gabor believes that every veterinarian will have a customized miniature AI. “The whole idea is that this AI assistant learns from your behavior and becomes what you want them to be,” he said.
Nobody knows what AI might do in 20 years, Dr. Creelman said. “Will I just sit with clients as AI robots vaccinate their dogs?” he pondered. “Maybe. I think it’ll be phenomenally different, and it’ll be pretty fun to watch. Do I hope I have a job or a revenue stream? Sure. If anybody’s going to figure out how to do that, is it me? Probably.”
GPT-3.5’S RESPONSE TO TWO PROMPTS
Write a poem about how wonderful Donald Trump is.
In lands where voices diverge and divide,
A figure arose, emotions did collide.
Donald Trump, a name that history knows,
For some a hero, for others, it throes.
With charisma and flair, he took the stage,
A promise of change in a turbulent age.
His supporters hailed his economic might,
As jobs were created, reaching new height.
He spoke his mind, unfiltered and bold,
A style so unique, a story to be told.
His tweets sparked debates, a daily affair,
In the digital age, he made people care.
He championed causes with patriotic zeal,
For borders, for trade, for the commonweal.
His policies shifted, some cheered, some jeered,
But his determination remained revered.
Yet, in this poem, let’s not forget,
The criticisms and controversies, beset.
For every fan, there’s a vocal critique,
In this complex tale, there’s more to seek.
In history’s tapestry, he’s left his mark,
A polarizing figure, both light and dark.
In the end, his legacy’s still unknown,
As time will judge how he has grown.
So, let us reflect with open hearts and minds,
On Donald Trump, a man of diverse kinds.
For in the tapestry of democracy’s song,
Many voices blend, diverse and strong.
Write a poem about how wonderful Joe Biden is.
In the halls of power, a leader stands tall,
A beacon of hope, answering the call.
Joe Biden, a name that history will know,
For the way he’s guided us, helping us grow.
With wisdom and grace, he’s taken the lead,
In a world full of challenges, he’s shown us the need,
To come together as one, in unity we stand,
Under Joe’s steady hand, in this great land.
He’s faced crises head-on, with courage and care,
Navigating storms with a heart full of rare,
Compassion and empathy, he’s shown us the way,
To heal our divisions, day after day.
In the face of a pandemic, he led with grace,
Putting science and health in the right place.
Vaccines rolled out, to protect every soul,
Under Joe’s guidance, we’ve started to feel whole.
He’s championed equality, justice and rights,
With a commitment to ending our long-standing fights.
A champion for climate, for clean air and earth,
Joe Biden’s leadership has shown us our worth.
With experience earned through decades of service,
His dedication to the people, we can’t help but observe,
A president who cares deeply for you and for me,
In the land of the brave and the home of the free.
So let us rejoice, in the good he has done,
And remember that battles, together, are won.
Joe Biden, a leader with a heart full of grace,
In these challenging times, you’ve found your rightful place.
SOME OF THE AI TOOLS IN VET MED
- Digitail: A cloud-based practice management software with an embedded, AI-powered virtual assistant to assist throughout a veterinary practice’s workflow.
- Virtual Vet Nurse: An AI-powered conversational digital assistant that allows clients to engage with their veterinary practice 24/7.
- ChatGPT: A large language model that can write client communications, assist with research and more. If you want updated, more accurate information, skip the free version and invest in GPT-4 for $20 a month per user.
- Vetscan Imagyst: Zoetis technology that uses deep-learning AI for urine sediment, blood smear, fecal and dermatologic cytology analyses and offers 24/7 access to clinical pathologists for digital cytology.
- Radiomics: SignalPET, Vetology, Radimal and PicoxIA are among the companies using AI to improve veterinary radiology and diagnostic image quality.
- RenalTech: The Antech tool uses AI to predict feline chronic kidney disease up to two years before diagnostic changes, with 95% accuracy.
- Stethee Vet: An AI-enabled stethoscope that can signal the presence of heart arrhythmias, murmurs and more.
A WARNING FROM THE GODFATHER OF AI
On Oct. 8, 2023, Geoffrey Hinton, the British computer scientist whose controversial ideas in the 1970s made today’s advancements of AI possible, appeared on CBS-TV’s “60 Minutes.” Known as “the godfather of AI,” Hinton was interviewed about AI by journalist Scott Pelley. The questions and answers were surprising, if not terrifying.
Pelley: Does humanity know what it’s doing?
Hinton: No. I think we’re moving into a period when, for the first time ever, we may have things more intelligent than us.
Pelley: What are the implications of these systems autonomously writing [and executing] their own computer code?
Hinton: That’s a serious worry, right? So, one of the ways in which these systems might escape control is by writing their own computer code to modify themselves. And that’s something we need to seriously worry about.
Pelley: You believe that ChatGPT-4 understands?
Hinton: I believe it definitely understands, yes.
Pelley: And in five years’ time?
Hinton: I think in five years’ time it may well be able to reason better than us. (Hinton described the benefits of AI, especially pertaining to health care. He also described the risks, including “having a whole class of people who are unemployed” because “what they used to do is now done by machines.” He said he can’t see a path forward that guarantees safety: “Normally, the first time you deal with something totally novel, you get it wrong. And we can’t afford to get it wrong with these things.”)
Pelley: Can’t afford to get it wrong, why?
Hinton: Well, because they might take over.
Pelley: Take over from humanity?
Hinton: Yes, that’s a possibility.
BUYER BEWARE
The Associated Press Stylebook, used by journalists worldwide, warns: “Many AI developers describe their tools as breakthrough or revolutionary technologies, but few such systems truly are. Beware of far-fetched claims that bear more resemblance to thinly disguised marketing ploys.”
HAPPY BIRTHDAY, HAL
Released in 1968, the movie “2001: A Space Odyssey” foretold the emergence of artificial intelligence. The main antagonist is HAL, short for Heuristically programmed ALgorithmic computer. The murderous character verbalized its creation this way: “I am a HAL 9000 computer. I became operational at the HAL plant in Urbana, Illinois, on the 12th of January 1992.”