Priorities and Recommendations for Using Artificial Intelligence (AI) to Improve Equid Health and Welfare.

Artificial Intelligence (AI) has the potential to enhance the health and welfare of equids, and this study gathered expert consensus on the most important areas for AI development and implementation to maximize benefits. It also identified major welfare concerns, barriers to AI adoption, and practical recommendations to overcome these challenges.

Study Objectives and Approach

• Aim: To establish consensus among stakeholders on how AI can be effectively developed and used to improve equid (horses, donkeys, mules) health and welfare.
• Method: A workshop involving 41 participants with expertise or interest in equid welfare was convened to generate statements regarding:
• Current welfare concerns in equids
• Priority areas where AI could have the greatest impact
• Barriers to the successful implementation of AI
• Solutions and recommendations for overcoming these barriers
• These statements were further refined and validated through Delphi surveys, requiring at least 75% agreement among respondents for a statement to be accepted.

Key Found Welfare Concerns

• Ethological needs of equids are often not met – meaning their natural behaviors, social interactions, and environmental requirements are inadequately addressed.
• Poor management practices contribute significantly to welfare issues.
• Insufficient knowledge and understanding among equid owners and carers about proper care and behavior of animals were identified as major factors affecting welfare negatively.

Priority Areas for AI Development

• Assessment of equid wellbeing through automated or AI-driven tools, including behavioral monitoring and health evaluation.
• Individual-level monitoring, such as tracking the health and behavior of single animals for early detection of welfare problems.
• Population-level monitoring, where AI can analyze data across groups of equids to identify trends, outbreaks of disease, or welfare issues.

Identified Barriers to AI Implementation

• Limited understanding of both equine behavior and AI technology among potential users and developers, leading to challenges in designing effective tools.
• Biased, unethical, or insufficient data collection hinders the development of reliable AI models.
• Difficulty in developing accurate AI models that generalize well to diverse real-world conditions.
• Challenges in validating AI systems to ensure they work as intended and provide trustworthy outputs.
• Uncertainty in interpreting AI-generated data, limiting user confidence and uptake.

Proposed Solutions to Overcome Barriers

• Development of AI systems that are evidence-based and unbiased, ensuring high scientific standards and ethical data practices.
• Adherence to best practice guidelines in AI development and application for equid welfare.
• Implementation of approval or regulatory frameworks to validate and monitor AI tools before widespread use.
• Encouragement of collaboration among AI developers, equine experts, and end-users to align technology with practical needs.
• Focus on education and training for AI users to improve understanding of both equine behavior and AI capabilities.

Significance and Future Use

• This is the first study to systematically gather stakeholder opinions regarding where AI can best benefit equid welfare, potential obstacles, and how to address them.
• The consensus findings provide clear guidance to prioritize research funding and development efforts focused on high-impact areas.
• The recommendations aim to accelerate the ethical and effective implementation of AI technologies to support equid health, welfare assessment, and management.