We are talking “chatbots” today! What are the possibilities and limitations within the veterinary world and is there an ethical dilemma?

You may have become aware of the hype around artificial language models or “chatbots” like Chat-GPT created by OpenAI, and the soon to be released google version, Bard. These systems are designed to produce human-like text responses to your questions using information from its internal programming. Chat GPT can’t currently access the internet, databases or search engines, and so can’t fact check, provide references, perform calculations or translations. But how far could this artificial intelligence (AI) technology go? And could it be useful or detrimental when used in the veterinary world?

Table of contents

What are the uses of technology in a veterinary practice?

Technology is creeping into a lot of aspects of our lives, and it is here to stay. Over the years we’ve seen technology aid in diagnosis and treatment of our pets, and the capabilities of that technology has developed. 

In your veterinary surgery, your veterinary team will already be used to using specialist equipment; such as laboratory blood and urine testing machines, ultrasound and x-ray machines, endoscopes; as well as monitoring equipment to check your pets’ safety during surgery or procedures. The take home here though, is that these pieces of technology are used alongside suitably qualified people, such as registered veterinary nurses or vets. The vets will assess information from lab results, imaging or monitoring equipment and make clinical decisions based on their expertise. Your pets’ welfare is at the forefront of your vets’ mind during this process. 

The veterinary field is embracing developing technology with instant chat and video calls, in the form of telehealth. We have seen new regulations developed in recent months, to allow modern communication with millennials; a style of pet health care that a lot of pet owners now prefer. The regulation developed aims to ensure telehealth supports the physical veterinary clinics without compromising the care given to pets and their families. 

Some veterinary companies, more so in the USA (according to the research I can find), already use AI technology, in interpreting abnormalities in x-rays. The AI is trained and learns from its interactions and from specialists in the field to improve its error rate in finding abnormalities in x-ray images. It is, however, absolutely not, a replacement for the specialists’ expertise. These technologies are programmed to spot x-ray abnormalities and not answer questions or interact with the user directly. 

How are “chatbots” different?

AI, such as Chat GPT, is different to currently used technologies, as it is essentially a computer interacting directly with a human user. This system depends on human input but relies on its programming entirely for output and responses. Human-like responses are created, that develop as the system “learns” according to the interaction. Humans should not take the responses as absolute fact but of that generated by a complex algorithm, with significant limitations. 

What are the capabilities of AI in veterinary medicine?

AI has the ability to search thousands of textbooks and journals simultaneously and produce summaries of those texts. This is a common application of natural language processing technology (NLP) carried out by AI technologies. Veterinary professionals could use text summaries to work up complex cases that ultimately, could help to treat your pet. 

AI can be taught to read x-rays or ultrasounds. Or it can be taught to aid in disease surveillance and help in predictive analytics; (identifying patterns and predicting which animals are at highest risk for certain health conditions). AI could be used as a tool for medical professionals. But I note that with respect to medical information, AI needs to be asked the right questions and have skilled professionals interpret that information to give accurate responses.

In the wrong hands, just like “Dr Google” as I call it, this rather clever tool could give inaccurate information or even be interpreted in a harmful way by the wrong – or simply misinformed – user. Listed on the “new questions” home page, OpenAI acknowledge that their system “may occasionally generate incorrect information” or “may occasionally produce harmful instructions or biased content” and that it has “limited knowledge of world events after 2021”. These are obviously significant limitations to an AI technology “chatbot”; especially at this new, early stage of its development where the systems programming and database is new and in its test phase. 

So, what are the limitations of AI?

Differing from the technology we use day to day in veterinary care, AI technology is designed to function independently of a human. The technology is not a replacement for veterinary care or specialists’ expertise. 

I actually posed this question to Chat-GPT, below is the answer created: 

“As an AI language model, I can provide general information on veterinary medical conditions based on my training and knowledge of the subject matter. However, I have several limitations when it comes to diagnosing veterinary medical conditions.

Firstly, I am not a licensed veterinarian and do not have the ability to physically examine animals. A physical examination is an essential component of the diagnostic process for any veterinary medical condition, and it is not something that I am capable of performing. Without a physical examination, it is difficult to accurately diagnose many medical conditions, as certain symptoms may be invisible or difficult to detect without specialized equipment.

Secondly, veterinary medicine is a rapidly evolving field, with new research and developments being made all the time. While I am continually updated with new information, my knowledge is only current up until my training data cut-off date, which was September 2021. Therefore, I may not have access to the latest research and developments in the field.

Lastly, every animal is unique, and many factors can influence their health, including age, breed, genetics, lifestyle, and environment. These factors can make it difficult to diagnose medical conditions, even for experienced veterinarians. As an AI language model, I do not have the ability to consider all of these factors when making a diagnosis, and I am not capable of individualizing diagnoses based on specific animal cases.” 

In short…

In short, diagnosing your pets’ ailments is a complex process which draws on a lot of skills. Your veterinary surgeon will be considering your pet’s signalment (species, breed, age, sex); along with any previous medical history and your own personal circumstances. All before making their diagnosis and developing a treatment plan. In order to stay current and up to date, your vet will be completing a mandatory number of hours of continuing professional development “CPD” each year to keep up to date on new developments in the veterinary field. Knowledge is always evolving and not a static process which as Chat GPT acknowledges, is outside of its capabilities. 

Perhaps most importantly, your vet, uniquely to being human, will also consider what is ethical. 

Should we consider the ethics of asking an AI veterinary questions?

What is right for your pet? One of the first ethical promises we make as veterinarians is to “first, do no harm”. We promise, in our declaration and our code of conduct, to “make animal health and welfare our first consideration when attending animals”. We also promise to “be open and honest with clients and respect their needs and requirements”.

AI is not human and can’t individualise or consider your pet’s circumstances as a member of your family. AI can’t consider your ethical or religious beliefs or financial constraints and work with you to get the best outcome for your pet whilst considering these complex factors. Your vet is able to make sometimes complex diagnoses AND consider the human factors. There is a reason we spend 5+ years studying to be able to care for the many complexities of your pets’ health. This can’t be replicated by a computer. The computer even acknowledges that! 

Should you trust an AI with vet questions? 

In short, no! I asked Chat GPT this question, and this is the reply I got: 

As an AI language model, I can provide you with accurate and relevant information based on what I have been trained on. However, it’s important to note that I am not a licensed veterinarian. Any information I provide should not be considered a substitute for professional veterinary advise, diagnosis, or treatment. 

If you have any concerns about the health of your pet, it is always best to consult with a licensed veterinarian who can examine your pet and provide a proper diagnosis. They have the necessary treating and expertise to help you and make informed decisions about your pets’ health and well-being. 

In developing this blog, I asked Chat GPT quite a few veterinary questions that the pet owner might ask, in order to play around with its capabilities. Some answers were plausible, and content medically correct, other times the answers were nonsensical or contained biased and incorrect content. It was an interesting experiment to confirm my suspicions that AI is a bit of fun but not a safe source of medical information for pet owners at this time. 

In summary, AI is here to stay

It is a useful tool in many areas of modern life, including veterinary medicine. AI may speed up some processes, freeing humans up to achieve more and help more people and their pets. However, AI has significant limitations, especially early on in its development. AI can only produce output that is within its programming capabilities. And although these appear very human-like, they lack the “human factor” of empathy and ethical decision making. AI is unable to replicate the complexities of medical decision making; which leads to the diagnosis and navigating development of a treatment plan with many individual patient factors to consider. 

Your vet is unique. AI won’t replace them just yet. 

The veterinary field is embracing technology. AI will be a part of our future as it will be in human medical communications. Who knows how far it will go. 

What I will emphasise though, is that if you have concerns over your pets’ health or welfare, there is not an AI computer model or even a community forum that will substitute for the care you will receive from a suitably qualified veterinary surgeon. Call your clinic for expert advice. 

Resources: