August 2024
In this VETgirl online veterinary continuing education blog, Dr. Sonja Olson and Dr. Liz Barton discuss how artificial intelligence (AI, not to be confused with veterinary artificial insemination!) might impact the 5 domains of animal welfare for better or worse – both within and beyond the clinic – and how and why we need to advocate for animals in this space. Please note the opinions in this blog are the expressed opinion of the authors, and not directly endorsed by VETgirl.
Artificial Intelligence and the Wellbeing of Animals – Is This Technology Friend or Foe to Our Animal Friends?
By Dr. Sonja Olson & Dr. Liz Barton
Introduction
In April, our first blog How Can AI Address Burnout in Vet Med examined the potential impacts of AI on veterinary professionals’ well-being. In particular, we focused on what we have come to understand about some of the psychosocial challenges that contribute to burnout in veterinary practices and how AI might mitigate, or exacerbate, the development of occupational distress associated with work-related burnout.(1) In this second blog, we wish to shift the focus towards the impact of AI on the wellbeing of the animals that are in our collective care. Veterinary professionals work collaboratively with other disciplines and stakeholders globally to advance One Health, “supporting healthy, productive, and resilient communities and ecosystems…including protecting and advancing the health and welfare of all animals.”(2) The intersection of animal-human health is one area where advancements in AI are happening apace and have the potential to hugely impact the wellbeing of sentient beings.
To raise awareness and stimulate reflection, the following statistics are valuable to consider when we reflect more broadly upon the animals whose welfare is impacted by One Health initiatives and the global animal caregiving community:
- 31.0 billion land animals and 38.3 to 215.9 billion fish are being farmed globally at any given time (3)
- As of 2022, there were over a billion pets estimated to be in households worldwide (4)
- There are approximately 58 million owned horses in the world (5)
- Worldwide, AZA-accredited zoos host approximately 800,000 animals (6)
Wellbeing in animals
Wellbeing is such a broad and complex term for both animals and humans. In many conversations around human wellness and supporting a flourishing life, the “Wellness Wheel” that examines 8 different realms is commonly utilized. In today’s blog, we will use the 2020 Five Domains Model of Animal Welfare (7) as an evidence-based guide on potential impacts of AI on different aspects of animals’ lives. Additionally, the 2020 updated model includes the ethical and philosophical considerations associated with a “life worth living.” For animals to have “lives worth living”, it is necessary, overall, to minimize their negative experiences and at the same time to provide the animals with opportunities to have positive experiences that support their holistic ‘wellbeing.’(8)
Overview of areas of impact of AI
When we think about how AI might impact animal wellbeing, there are two broad areas to consider: interpretive (i.e., data processing and intelligence) and generative (i.e., creative) AI.
Interpretive uses focus on the practical application of data and technology to enhance animal health directly. These include areas such as:
- Epidemiology: Using data to forecast health trends and outcomes, enabling proactive management of animal health.
- Monitoring and Diagnostics: taking data from wearables, imaging and lab diagnostics to improve speed and accuracy of diagnosis.
- Treatment: assisting veterinarians in making informed decisions about treatment based on a variety of data, including genetics, clinical parameters, signalment and history.
Generative AI includes image, audio, and text creation, which may influence animal wellbeing directly and indirectly:
- Content creation: Educational content explaining animal health issues, appropriate husbandry, and the importance of veterinary check-ups, helps disseminate information on animal health topics to animal owners. While this use of AI does not directly influence animal health, it could play a pivotal role in communication and awareness of good care.
- Animal communications: In several species, AI is being used to ‘decode’ audio and visual communications in animals.
Generative models can subsequently be used to generate lookalike communications, for example the CETI project is currently decoding whale song with the aim to be able to communicate back to the whales. While this ‘Dr. Dolittle’ potential may have huge benefits such as stimulation, understanding, and warning of danger (e.g. to reduce risk of entrapment in fishing nets), the impact on the mental wellbeing, behavior and cultures of these animals is unknown, presenting an ethical dilemma. Just because we will soon be able to ‘talk to the animals’, should we, and how do we ensure this ability is not abused?
5 Animal Welfare Domains
Nutrition: The complexities of diet constituents and the plethora of individual factors that affect how nutrients are processed makes this an ideal area for AI to improve our understanding. By quickly analyzing large data sets, AI can provide a range of information, from the health status of animals to their feeding habits. We can therefore tailor nutrition more precisely in accordance with the nutritional needs of the animals.
For example, models already exist that can identify cats at risk from renal disease up to two years prior to the onset of clinical changes, enabling earlier dietary intervention. In farm animals, AI is being used to analyze the DNA of the gut microbiome of cattle herds, and subsequently identify food additives to reduce methane outputs from cows. While the primary benefit here is to reduce greenhouse gas emissions, such uses could expand to optimizing health for the animal.
Environment: Through constant monitoring of environmental parameters, such as temperature, light, ventilation, and toxin levels, and the ability to automate building controls, AI can help to optimize the environment to improve health and wellbeing. In the open, geotracking of ranging animals can aid retrieval of individuals that become trapped or separated. This technology is also being used by conservationists to identify at-risk areas and prioritize resource allocation to protect endangered species (e.g., by combining data that predicts risk of wildfires with animal population data to plan proactive interventions).
Health: Earlier detection of disease, injury, or functional impairment by AI has been available for years and is progressing quickly. Biometric profiles and variations in postural behaviors for companion animals, cattle and sheep allow for earlier assessment, diagnostic evaluation, and therapy for a wide variety of health concerns. Technology to assess pain, (e.g., pain monitoring of horses through mapping human facial expressions of pain or rapid-detection tool that uses AI to ID early variation to postural behaviors that may signify lameness), supports more rapid evaluation and therapeutic interventions to be provided.
These AI-driven tools combined with tele-communication with veterinary teams including language translation technology when appropriate increases the likelihood of animal guardian understanding and compliance. Other health-care apps that are being developed such as TT Care allows companion animal owners to upload pictures of their pet (e.g. eyes, skin, teeth), and per the website, receive an AI analysis of their pet’s possible health condition within minutes, claiming a 93% accuracy rate. The benefit of such apps is that it can serve to inform and educate a concerned pet owner and then lead to the appropriate veterinary consultation where more thorough examination and diagnostics can occur. Maintaining the veterinary-client-patient relationship when using these apps will be essential for ethical and appropriate pet care.
In all of the above instances, there is the capacity for increasing health literacy, as well as more efficient and informative communication with veterinary care providers. The human-animal bond is strengthened when expedient support of animal physical health and addressing of associated pain is provided. In the case of farm animals, disease reduction and increased productivity benefits the animals, the humans involved in the animal care, and supports early detection of epidemiological trends allowing for earlier intervention (important on so many levels!).
Image by Gordon Johnson from Pixabay
Behavior: This is a fascinating and evolving domain in animal welfare as it includes animals exercising agency when interacting with their environment, with other non-human animals, and with human beings. The positive or negative holistic impacts and subsequent responses of an animal to situation-related factors is the focus in this domain. AI-driven systems to monitor and to learn from animal behavior are being used with pets as well as on livestock farms. The systems can improve environmental conditions for animals and can alert caregivers if there are concerning aberrations.
Engaging with the monitoring and learning from an animal’s behavior was taken in a different direction with a new “robotic dog nanny”, created by the company ORo. This cyberfriend was designed to meet the needs of companion animals when they are at home alone. The technology creators state that their desire is to address the pet’s nutritional, physical, mental and emotional engagement needs when owners are not physically available to engage with the pet. Devices such as this advanced robotic companion and cameras that monitor and feed and/or play with cats and dogs (e.g., PetCube) can bring peace of mind to owners, can provide safety for the animals, and possibly increase environmental enrichment for the pet. However, animal behavior researchers and ethicists raise the reasonable concern that beyond potentially compromising privacy, these technologies may have unforeseen negative consequences on the cherished human-animal bond (9).
Other AI-related technologies that would fall into this behavioral domain would be the myriad of tools being developed to decode animal communications. There is a desire by many to better understand and to thus enhance the relationship between human and animal. This is the case for the creators of MeowTalk (an AI-powered cat translator app) and Zoolingua (a dog translator app). Will understanding animals’ emotions, including our pets, enrich our relationship with them? Will these tools inadvertently cause miscommunication or misinterpretation with animals causing harm? In the case of wildlife such as whales as was raised earlier in this blog, what are the ethics of mimicry in interfering with their natural environment and inter-species bonds?
Mental domain: This fifth domain of animal welfare has come to acknowledge how both biological functioning impacts and affective experiences importantly contribute to an animal’s subjective experience and holistic wellbeing. This brings us back to the concept of a ‘life worth living’. The subjective experiences of animals might include fear, pain, anxiety, distress as well as satisfaction, security, environmental enrichment, and meaningful companionship. Research and understanding of the integral elements for animal health and wellbeing have the potential to be advanced by and supported by the development of AI-driven technology. However, a willingness and humility to bring interdisciplinary teams together to offer ethical and scientifically-sound experience will be essential. Balancing the human-centric benefits of all of these technologies and deeply considering the impacts on the 5 domains of animal welfare is paramount to truly support the One Health mission and perspective.
Conclusion
There are many and varied areas where AI has the potential to impact animal health across the fields of veterinary, conservation and livestock farming. However, whether the outcomes confer a positive or negative impact on animal wellbeing is in the balance. As veterinary professionals, we are the ones charged with safeguarding animal health and welfare. It is incumbent on us to take the lead on this. In truth, the potential to impact animal wellbeing through AI is greater than through any other aspect of veterinary medicine.
We need veterinarians and veterinary technicians to be involved and advocating for animals as this technology is developed. We hope this article inspires you to think about what you can do within your practice to ensure AI works for the benefit of animals in our care, and beyond to the billions of animals on the planet who have no choice in this and need us to be a voice for the voiceless (until GenAI does that too!).
Cited Resources
1. https://vetgirlontherun.com/how-can-ai-address-burnout-in-vet-med-by-dr-sonja-olson-dr-liz-barton-vetgirl-veterinary-continuing-education-blog/ (April 2023)
2. AVMA website, https://www.avma.org/resources-tools/one-health
3. https://ourworldindata.org/how-many-animals-are-factory-farmed (September 2023)
4. https://healthforanimals.org/reports/pet-care-report/global-trends-in-the-pet-population/ (September 2022)
5. https://worldmetrics.org/horse-statistics/ (June 2024)
6. https://www.aza.org/connect-stories/stories/interesting-zoo-aquarium-statistics (May 2021)
7. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7602120/ (September 2020)
8. https://www.mdpi.com/2076-2615/6/3/21 (March 2016)
9. https://www.washingtonpost.com/business/interactive/2024/smart-pet-technology-care-ai/ (June 2024)
Please note the opinions in this blog are the expressed opinion of the authors, and not directly endorsed by VETgirl.
Only VETgirl members can leave comments. Sign In or Join VETgirl now!