Science

AI study flags welfare gaps in greyhound racing data

Researchers used AI agents to analyze UK greyhound racing records and found rising on-track fatality rates and limited public data on dogs’ welfare.

Lucas Ferreira

By Lucas Ferreira · Science & Environment Writer

3 min read

AI study flags welfare gaps in greyhound racing data
Photo: Phys.org

University of Melbourne researchers say an AI-assisted review of UK greyhound racing records has exposed welfare concerns that public industry summaries did not clearly show. The findings add to scrutiny of a sport already facing bans and phaseouts in several jurisdictions, according to the university.

The research team said it used AI agents to compile public data on 31,028 licensed greyhounds and more than 1.26 million race starts in the United Kingdom from January 2022 to March 2026. The university described the method as agentic AI: software used for repetitive tasks while remaining under human supervision.

Lead researcher Dr. Mia Cobb, an animal welfare scientist at the University of Melbourne, said the system brought together information from several public websites at a scale that would otherwise have required months of work by a research team. Cobb said the results showed the limits of public reporting by the UK racing industry.

According to Cobb, the industry had reported a stable on-track fatality rate even as the number of deaths rose and the number of races fell. Further analysis found the on-track fatality rate for racing greyhounds increased by 30% between 2022 and 2024, she said.

Cobb said the way the regulator rounded fatality rates to two decimal places over the past three years hid the pattern of more dogs dying while race starts declined. She also said more than two dogs had died on track each week since the UK greyhound welfare strategy began in 2022.

Dr. Simon Coghlan, a co-researcher and deputy director of the University of Melbourne’s Centre for AI and Digital Ethics, said the public record remains incomplete. Coghlan said public registries do not account for the welfare of individual dogs after racing, including deaths that may follow injuries sustained on the track.

Coghlan said the Greyhound Board of Great Britain holds unpublished information on where dogs go after retirement, why careers end, and individual-level injury and euthanasia figures linked to specific racetracks. He said keeping that information from public view limits accountability for welfare claims.

The study also found high turnover in the racing population, according to the researchers. Cobb said about 40% of racing dogs stop each year, while the industry maintains numbers through a steady intake of new dogs from Ireland.

According to Cobb, Irish-bred dogs account for 85% of greyhounds racing in the UK, which she said represents nearly half of the dogs bred annually in the Republic of Ireland. The researchers also reported that the typical greyhound races for less than a year.

The team said its analysis identified differences between tracks in the rate of harmful racing events, including crashes and falls. Cobb said those track-level details had not been disclosed by the regulator, including during questioning by a Welsh parliamentary committee considering a ban.

The research comes as governments in several places reconsider greyhound racing. The university said Scotland and Wales announced plans earlier this year to ban or phase out the sport, while New Zealand announced in December 2024 that racing would be phased out and prohibited by July 31 this year.

In Australia, greyhound racing is already banned in the Australian Capital Territory, according to the university. Tasmania plans to phase it out by mid-2029, South Australia’s industry faces a July 2026 reform deadline or a possible statewide ban, and Western Australia’s Parliament is holding a formal inquiry.

RSPCA Australia Chief Science Officer Dr. Suzie Fowler said the Australian industry has long faced welfare problems, including overbreeding, poor living conditions, injury and death risks, and limited transparency about dogs after retirement. Fowler said AI tools could help identify welfare problems and assess reforms, but their value depends on what data the industry makes public.

Cobb said the researchers hope the work helps policymakers assessing greyhound racing rules. She said public disclosure does not necessarily show how dogs are faring and called for greater accountability in industries that rely on animals.

This story draws on original reporting from Phys.org.