BY: SAMANTHA BARTLETT, DVM

Artificial Intelligence (AI) has become a hot topic in the world lately, especially on university campuses. AI technology involves machines simulating human intelligence based on a database of patterns. Despite the relative novelty of the technology, AI is already in use in veterinary medicine. 

The AI being used in veterinary medicine is classified as a narrow-use application, meaning it only does one thing. AI is being used to analyze research data and predict outcomes in veterinary research programs. For example, diagnostic x-rays are fed into the system to get an output on disease classification. This program only does one specialty and works with a specific set of parameters to give information. It is attractive because it can increase the efficiency and speed of the veterinary practitioner.

In contrast, general AI can take in multiple information sources and analyze them to produce diagnoses or treatment plans. While this application can be attractive in a clinical setting, it does have several regulatory and ethical drawbacks. The application of general AI to medical diagnostics involves combining input data, imaging data, and genomic data. Somehow, these programs need to be able to integrate the long-term experience of qualified clinicians for decision making. Currently, these models are not as reliably effective as human veterinary professionals. 

While AI applications in healthcare are moving forward at a rapid pace, trust and ethical standards still need to be addressed. Currently there are no regulations for the use of artificial intelligence in medical diagnosis. Frameworks need to be established before widespread AI can be implemented in the veterinary medical field. For example, if AI makes a wrong prognosis resulting in an owner euthanizing their dog, there is no clear liability for the error. For now, it may be best to use AI as a diagnostic reinforcement until adequate testing and regulation have evolved.