RG Benchmarks is the first flagship project under OpenRG, focused on developing the industry’s first sector-led benchmarking framework for AI algorithms used to identify at-risk players. Despite the growing use of machine learning to support responsible gambling, there is currently no standardized way to evaluate the fairness, accuracy, and effectiveness of AI-based systems. RG Benchmarks addresses this gap by introducing a transparent, objective evaluation framework that enables meaningful comparison of risk detection models over time and across providers. By supporting researchers, operators, regulators, and vendors, this initiative aims to drive innovation, improve harm prevention efforts, and raise global standards for AI accountability in gambling.
We invite you to review our full preprint paper where we examine the current landscape and our vision for a standardized evaluation methodology: Read the paper