University of Melbourne researchers have unveiled troubling insights into the welfare of greyhounds in the United Kingdom’s greyhound racing industry, reigniting an international debate on whether the sport should face outright bans globally. This groundbreaking study, published in Frontiers in Animal Welfare and Policy, leverages cutting-edge artificial intelligence to analyze an unprecedented volume of data, exposing critical issues previously obscured by industry reports and regulatory opacity.
For the first time, AI agents were employed to aggregate and analyze data on over 31,000 licensed greyhounds, spanning more than 1.26 million race starts across the UK between 2022 and 2026. This massive dataset was drawn from publicly accessible online registries, but the innovative use of AI allowed for the rapid mining and synthesis of information that would have taken human researchers months to process. The research team, led by Dr. Mia Cobb from the Melbourne Veterinary School, demonstrated how these AI tools can peel back layers of industry obfuscation, presenting a more accurate portrait of greyhound welfare.
The greyhound racing industry has long marketed itself as transparent and accountable, routinely publishing annual welfare figures. However, Dr. Cobb’s analysis reveals these published figures mask a grim reality. Initial observations showed a seeming paradox: while the number of races declined, reported on-track fatalities stayed stable. Upon deeper investigation, it became apparent that the fatality rate had surged by 30% from 2022 to 2024, a critical statistic hidden by the regulator’s practice of rounding fatality figures to just two decimal places.
This failure in precise data representation has effectively concealed an alarming increase in dog deaths on the tracks. Since the implementation of the UK’s greyhound welfare strategy in 2022, more than two greyhounds have perished weekly in racing incidents. This stark finding challenges the official narrative of improving canine safety and underscores the urgent need for better data transparency and regulatory oversight.
Co-author Dr. Simon Coghlan, deputy director of the University of Melbourne’s Centre for AI and Digital Ethics, warns that publicly reported fatalities likely represent just a fraction of the actual welfare issues. Public registries do not track injuries sustained during races that lead to deaths or euthanasia after the race finishes. Furthermore, data on injury severity, retirement outcomes, and euthanasia linked to specific racetracks remain withheld by the Greyhound Board of Great Britain, severely limiting accountability.
The study also exposed consistent high turnover within the racing greyhound population, with about 40% of dogs ceasing to race each year. This turnover is sustained largely through imports, with 85% of UK racing greyhounds sourced from Ireland, accounting for almost half of the dogs bred in Ireland annually. These dynamics highlight systemic pressures that contribute to welfare decline through overbreeding and constant replenishment of the racing stock.
In addition to population-level trends, the analysis revealed disparities in injury risk across different racing venues. Some tracks were associated with significantly higher rates of falls and crashes, information that has never been disclosed or discussed openly by regulators, despite parliamentary inquiries, such as those from the Welsh government debating a ban on the sport. The lack of track-by-track injury reporting conceals important contextual factors critical to understanding and mitigating risk.
The timing of this research is particularly poignant, given recent moves by several governments toward greyhound racing bans or phase-outs. Scotland and Wales have signaled intentions to prohibit the sport, New Zealand is on course to ban racing by mid-2026, and Australian states like the ACT and Tasmania have already enacted or are planning phase-outs. Multiple Australian jurisdictions are actively reviewing the future of the sport amidst burgeoning public and scientific pressure.
Dr. Suzie Fowler, Chief Science Officer at RSPCA Australia, contextualizes the findings within decades-long concerns about greyhound welfare, including rampant overbreeding, substandard living conditions, high injury and fatality rates, and opaque post-retirement fates. She underscores the crucial role of enhanced transparency and data availability to enable meaningful welfare improvements and to hold the industry accountable.
This pioneering use of AI agents demonstrates a scalable, efficient methodology for analyzing complex datasets from disparate public sources, opening a new frontier for animal welfare research. While the tools significantly lighten the heavy lift of data compilation, their efficacy remains bounded by the quality and transparency of publicly available information. Nevertheless, they offer a promising avenue for researchers, regulators, and journalists striving to illuminate industry practices and advocate for vulnerable animals.
The implications of this research extend far beyond greyhound racing. The novel methodology can be adapted to assess welfare issues in other animal-reliant industries, offering a powerful instrument to elevate ethical standards and inform policy globally. The researchers hope their work motivates greater openness and accountability, providing the public and policymakers with authentic visibility into the lived experiences of animals subjected to commercial exploitation.
Dr. Mia Cobb concludes with a call for authentic industry transparency — data disclosure that genuinely reflects the on-the-ground reality of racing greyhounds rather than sanitized narratives. Only through robust, verifiable data can industries be held to account and ethical reforms enacted in line with evolving societal values around animal welfare.
For full technical details and comprehensive data analysis, readers are invited to consult the original research article published in Frontiers in Animal Welfare and Policy, which is publicly accessible via the journal’s website.
Subject of Research: Animals
Article Title: Using AI agents to assemble population-level data for visibility and animal welfare insights: a case study of greyhound racing in the UK
News Publication Date: 2026
Web References: https://www.frontiersin.org/journals/animal-science/articles/10.3389/fanim.2026.1868726/
References: 10.3389/fanim.2026.1868726
Keywords: Animal rights, Artificial intelligence, Animal welfare, Greyhound racing, Data transparency, AI agents, Computational modeling
