AI to monitor 60 risk signals in Spain’s new online gambling system

• DGOJ opens consultation on AI model tracking 60 risk indicators across online betting and casino
• System uses XGBoost to analyse real-time behaviour and generate operator interventions
• Lotteries excluded, raising coverage concerns over SELAE and ONCE products

Spain’s Directorate General for the Regulation of Gambling (DGOJ) plans to deploy an AI-driven oversight system that monitors more than 60 behavioural and transactional signals to detect risky gambling and prompt early interventions.

A public consultation on the initiative is open until 25 September 2025. The initiative focuses on online sports betting, poker and casino, positioning Spain to implement one of Europe’s most comprehensive real-time player protection frameworks if fully adopted.

The project traces back to Spain’s safer gambling agenda and will rely on machine learning—specifically the XGBoost algorithm—to evaluate session frequency and intensity, deposit and withdrawal patterns, and behavioural shifts after wins or losses, among other signals, to assess potential harm in real time.

The regulator highlights live betting as a higher-risk environment due to its fast cadence, with consecutive days of play and cash-out habits among the indicators to be monitored under the model.

The DGOJ has framed the consultation as a collaborative process involving operators, technology firms, researchers, and citizens, seeking feedback on technical and legal design elements before implementation details are finalised.

Officials and trade coverage suggest the framework could become a benchmark in the EU if it delivers consistent detection and actionable risk scores across licensed platforms. The system will not cover lottery products provided by SELAE and ONCE, which remain the most popular form of gambling in Spain, creating a gap in overall coverage and drawing questions about completeness of harm prevention.

Operationally, the design points toward real-time alerts and tailored responses rather than blanket measures, enabling operators to deploy proportionate interventions such as time-outs, deposit limits, or direct contacts based on individual risk profiles generated by the AI.

The regulator has not yet detailed standardised intervention protocols or a timeline for pilots and rollout, leaving open questions on enforcement and consistency across operators and regions.

“It would be one of the most ambitious oversights introduced by a gambling regulator in Europe, but there is a long way to go before this happens,” said the DGOJ’s consultation summary.

“The purpose of the consultation is to help establish a new AI-driven system that flags 60 criteria that could be indicative of gambling-related harm and excessive spending. The surveillance system will use the XGBoost machine learning algorithm to assess multiple variables including session frequency and intensity.”

If successfully implemented, Spain’s model could shift compliance from reactive checks to proactive, data-led safeguarding, with AI scoring intended to catch harmful patterns earlier and guide timely operator action.