ANJ-Funded Study Finds 39% of French Teenagers Play Gambling-Style Games but Cannot Prove a Gateway

A survey of 5,000 French 15 to 17-year-olds funded by the Autorité Nationale des Jeux finds that teenagers who play both gambling-style video games and real-money games score far worse on problem gambling than those who only gamble. It shows an association and a sequence, not a cause, and its authors say so.

• The ENJEU-Mineurs 2025 study by ARPEJ, financed by the Autorité Nationale des Jeux (ANJ) and published on 19 October, found 39.0% of 15 to 17-year-olds had played at least one "hybrid digital game" with gambling-like mechanics
• Casino scenes inside video games such as Grand Theft Auto were the most common form, at 26.3%, ahead of free demo casino games at 18.1%, social casinos at 16.6% and loot box games at 13.0%
• Among teenagers who also gamble for money, 26.0% scored as high severity on the adolescent problem gambling scale, against 3.9% of those who only gamble, and 67.4% said they had played gambling-like games first
• The authors say the findings are "compatible with" a possible gateway effect but that the causal mechanism remains imprecise, and the data are self-reported from an online panel
• The ANJ announced no new measure; its president tied the findings to its thinking on JONUM, the tradable digital item games it now supervises, while Spain's loot box ban for minors lapsed this month

France Measures Gambling-Style Gaming Among Minors for the First Time

The Association de Recherche et Prévention sur les Excès du Jeu (ARPEJ) asked 5,000 French teenagers aged 15 to 17 about six kinds of "jeux numériques hybrides" (JNH): social casino apps such as CrownCoins and McLuck, free demo casino games, video games with gambling features such as GTA, loot box games such as Counter-Strike 2 and Ultimate Team, fantasy sports such as Mon Petit Gazon and Sorare, and Web3 games. The ANJ published the results on 19 October with a 39-page report by Marie-Line Tovar, Pauline Saliot and Jean-Michel Costes.

The survey ran online from 29 August to 16 September 2025 among teenagers recruited from family panels with The BVA Family, with parental consent, quotas by sex, age and region, and weighting to national statistics, according to the report. Four focus groups followed. The ANJ financed the study and its staff reviewed the report, which says its conclusions are the authors' alone.

In all, 39.0% of respondents, 1,958 teenagers, had played at least one JNH: 42.4% of boys and 35.4% of girls. The reference period is stated inconsistently: the report and summary say the previous three months, while the ANJ's release and the report's methodology note say 12 months.

Spending, Advertising and the Overlap With Real-Money Gambling

Most of this play is free. According to the report, 39.9% of JNH players had spent money on them, equivalent to 15.6% of all respondents. Among spenders, 19.4% of social casino players put in more than €50 a month, while 38.7% of loot box buyers spent less than €10. Of JNH players, 68.9% said they had seen adverts urging them to gamble for money, TikTok most often.

Gambling is illegal for under-18s in France, yet 42.6% of the sample said they had gambled in the previous 12 months, up from 34.8% in 2021, according to ARPEJ's February results. Some 1,151 teenagers, 23.0% of the sample, did both. On the Canadian Adolescent Gambling Inventory's severity scale, 26.0% of these "mixed" players scored high severity against 3.9% of gambling-only players. Among mixed players who pay for JNH, the figure was 40.7%, against 14.1% for those who play free.

The gateway finding rests on one retrospective question to mixed players: had they played games imitating gambling before gambling? Of the 1,169 who answered, 67.4% said yes, most often casino scenes in video games (42.5%); 10.4% named loot box games. On a gaming disorder scale, 7.4% of JNH spenders showed a severe disorder against 1.9% of non-spenders.

ARPEJ recommends better information for parents about microtransactions, teaching teenagers about publishers' business models, and that regulators, the games industry and civil society consider age verification and parental consent for spending in all JNH. ANJ president Pascal Chèvremont said the study documents the growing "gamblification" of games popular with minors and would inform thinking on Web3 offers such as JONUM under the SREN digital law's experiment. No new measure was announced.

Association Is Not a Gateway, and the Report Says So

The 26.0% against 3.9% comparison is striking, but the two groups differ in more than JNH play. Teenagers who gamble and also play casino-style games are, by construction, more engaged in gambling-like activity of every kind, and the published tables do not control for impulsivity or the number of gambling forms played. The 67.4% figure shows order, not influence: free casino-style games are so widespread that many gamblers would have played one first even if the games had no effect. The authors write that the causal mechanism "remains imprecise" and warn that online panels tend to overstate excessive behaviour, and some subgroup figures rest on small numbers, such as 64 mixed players who mainly play slots. Coverage presenting games like GTA as a proven gateway goes further than the authors do.

What It Does Show Is Spending, and Spending Is Where Regulators Can Act

The most defensible signal is money. Within the same group, teenagers who pay inside JNH score markedly worse than those who play free, on both scales. That matches the international research the report reviews, and it points policy at payment mechanics rather than casino imagery, which explains the modest recommendation: age checks and parental consent for purchases, not a ban. Belgium shows the limits of a ban: the report cites research finding that in mid-2022, 82 of the top 100 grossing iPhone games on the Belgian App Store still sold loot boxes despite the prohibition.

France Has a Regulator Without Jurisdiction Over Most of This

Under French law a loot box is gambling only if four criteria are met case by case, including an expectation of gain, so most of what the study measures sits outside the ANJ's powers. What it does supervise is JONUM, closed to minors, where Sorare, Sorare Pro, Meta Football League and CyLimit are listed as declared offers, and that is where Chèvremont pointed. Elsewhere the route has stalled: Britain chose an industry-led approach in 2022, the Dutch Council of State overturned a penalty on Electronic Arts over FIFA packs in March 2022, and Spain's ban lapsed with the Cortes. The study adds French evidence to the ANJ's growing focus on risky play, but turning it into rules for video games would need Parliament, not the regulator.

The ANJ now has French data showing that teenagers who mix gambling-style games with real gambling, and who pay for them, are a high-risk group. What it does not have is proof that the games create the gamblers, or the powers to regulate most of those games.