The Autorité Nationale des Jeux (ANJ) has placed fresh pressure on operators in France by presenting a new algorithm for detecting excessive gambling. The algorithm identified about 600,000 account-based players with a high probability of excessive gambling in the second half of 2025. According to ANJ, this accounted for 8.7% of all the players assessed by the algorithm.
Operators Face a Bigger Detection Target
The revenue figure shows why ANJ sees the issue as urgent. As mentioned by ANJ, those players generated €1.2 billion in gross gaming revenue, equal to 60% of total account-based GGR in the scope analysed.
ANJ also distinguished a separate, higher-risk category within this overall figure. Approximately 300,000 gamblers have been identified as “manifestly excessive.” ANJ urged operators to prioritize identifying this category of users.
Current Systems Still Miss Many Cases
French operators have improved their detection procedures, yet according to ANJ, it is still insufficient. Operators detected 89,000 excessive gamblers in 2025 compared to 31,000 in 2024, which represents a significant increase. However, the number still fell significantly short of the regulator’s estimation based on its algorithms.
ANJ framed the findings as both a player protection issue and a compliance benchmark for licensed operators. Operators are bound by law to detect and provide assistance to excessive gamblers. Support can include phone contact, tailored play limits, referral to support groups or treatment centres, and, in stronger cases, account closure.
The French gambling market continues to grow. According to data released by ANJ in April, total market GGR in 2025 rose to €14.1 billion, marking an increase of 3% over 2024.
How the ANJ Model Works
The algorithm was built based on account-level data transmitted regularly to ANJ by licensed online operators, FDJ and PMU account-based gambling services. The model uses 23 risk factors, among which are:
-
Financial movements;
-
Play frequency;
-
Account activity;
-
Use of moderation tools;
-
Player history.
Each player receives a score that places them into one of four categories: recreational, at risk, excessive, or manifestly excessive.
According to ANJ, the tool was tested against the Canadian Problem Gambling Index and evaluated by a scientific panel. Similar initiatives are ongoing in Spain and the Netherlands, although ANJ claims that its algorithm is the first of its kind in Europe.
What Comes Next
Algorithm usage by operators is optional. ANJ encourages companies to adopt the algorithm along with their other systems to help ensure compliance.
The authority will also use the tool for oversight. During the 2027 evaluation of the prevention action plan, ANJ could assess the operator data against the data from the algorithm.
French operators can no longer treat problem player detection as a single metric within the company. ANJ now has its own standard that is difficult for operators to justify when comparing their figures to the regulator’s estimates.
