60% of the 439 mobile games examined in a 2025 international sweep sent notifications nagging players to reopen the app, according to the International Consumer Protection and Enforcement Network's Mobile and Online Games Sweep, published 30 June 2025. Nearly half of the games swept also built in randomized loot box mechanics, and most did not disclose that up front. Below is what the largest recent app audits, and the biggest academic app-level studies, actually found, with every figure traced to its named source.
How many mobile games use dark patterns?
60% of the 439 mobile and online games examined by 22 consumer protection agencies sent notifications nagging players to log back in and claim a reward, and 32% pressured players into buying an item, a life, an energy refill, or a booster, immediately after they lost a level, according to the ICPEN Mobile and Online Games Sweep, published 30 June 2025. Urgency tactics, including countdown timers and low-stock warnings that were sometimes found to be false, appeared in 24% of the games, and 38% advertised offers as available for a limited time only.
Figure 1: Four manipulative design techniques found across the 439 mobile and online games ICPEN's member agencies audited. Source: ICPEN Mobile and Online Games Sweep, June 2025.
These are recent additions to a well-documented pattern. A broader 2024 sweep of subscription sellers, covering both apps and websites, found comparable prevalence across the wider consumer web; see our dark patterns statistics for 2026 for the fines and enforcement side of that story. None of the four game-specific tactics above were rare exceptions: each showed up in roughly a quarter to three-fifths of the games examined.
How common are loot boxes and hidden costs in mobile apps?
78% of the games swept had some form of virtual currency built in, and of those, 36% sold currency bundles that did not align with the in-game items players actually bought most, a setup that tends to leave players holding leftover currency they can only spend on more purchases. Only 2% let players cash out virtual currency for real money, and only 7% let players track their own spending history inside the app.
Figure 2: The gap between what ICPEN's auditors found inside the games and what each game's own store listing disclosed. Source: ICPEN Mobile and Online Games Sweep, June 2025.
Loot boxes, randomized rewards players can buy with real or virtual currency, were observed in 46% of the games swept, but only 14% of all games disclosed a loot box mechanic on their platform page. That means only 30% of the games that actually contained loot boxes told players about it up front. Of the games found to include loot boxes, 62% carried an age rating suitable for children, and in most of those cases the presence of loot boxes was not reflected in that rating at all.
Figure 3: The disclosure test ICPEN's auditors applied to each game's loot box mechanics. Source: ICPEN Mobile and Online Games Sweep, June 2025.
A loot box or an unlimited-feeling virtual currency wallet built into a game is far more likely to go undisclosed than to be flagged plainly on the app's own store listing.
Do dark patterns differ between iOS and Android apps?
The same 143 mobile apps, reviewed on both iOS and Android by researchers publishing in Behaviour and Information Technology in January 2025, carried more dark pattern instances on iOS. Across those 143 apps, iOS versions accounted for 1,477 total instances versus 1,398 on Android, a statistically significant gap the study's authors linked to differences in app store guidelines, developer tooling, and a rise in A/B-testing-driven design changes on each platform. Spread across 143 apps, that works out to roughly 10.3 dark pattern instances per app on iOS and 9.8 per app on Android.
Figure 4: Dark pattern instance counts across the same 143 mobile apps, reviewed on both platforms. Source: Li, Flatla, and Arndt, Behaviour and Information Technology (2025).
iOS's reputation as the more locked-down, safer platform does not extend to dark patterns. The researchers noted that this may leave iOS users less vigilant toward app-related risks than Android users, precisely because that reputation exists.
What do the largest academic app studies find?
The largest app-level percentage on record comes from an older study. Researchers at the University of Zurich analyzed 240 popular apps crawled from the US Google Play Store in 2019 and found that 95% contained at least one dark pattern, averaging seven distinct types per app, with false hierarchy present in 61% of apps and preselection in 60%, published at the CHI 2020 conference. No comparably large, comparably clean app-level percentage has been published since, which is itself worth stating plainly rather than papering over with an unrelated newer number.
Two more recent academic projects built larger app datasets but reported their findings differently, as detection-model performance rather than a simple prevalence percentage. UIGuard, published in 2023, assembled the largest academic mobile dark-pattern dataset to date, drawing from 1,023 apps and flagging 1,660 distinct dark-pattern instances of 16 types. AppRay, a follow-up detection system with a version dated April 2026, analyzed 100 apps across 22 categories and flagged 2,185 instances of 16 types across 876 of the 19,722 UI screens its automated exploration tool captured, roughly 4.4% of screens examined. The lower screen-level share reflects a stricter automated classifier and a per-screen rather than per-app measure, not dark patterns becoming rarer in newer apps.
How is dark pattern use regulated for apps that sell subscriptions?
A companion 2024 sweep by ICPEN, run alongside the Global Privacy Enforcement Network, examined 642 traders selling subscriptions through websites, apps, or both. 75.7% used at least one dark pattern and 66.82% used multiple, leaving only 156 traders, about a quarter of those examined, with none found at all.
Figure 5: Only about a quarter of the 642 subscription sellers examined were found to use no dark patterns at all. Source: ICPEN and GPEN Dark Patterns in Subscription Services Sweep, July 2024.
Among traders offering auto-renewing subscriptions, 81% gave consumers no way to turn off auto-renewal inside the purchase flow itself, 70% gave no information on how to cancel during enrollment, and 67% gave no clear date by which a consumer needed to cancel before being charged again. Among traders that required payment details before starting a free trial, that requirement showed up in at least 66.4% of the cases investigated.
For any business selling subscriptions through an app, most of that list is a disclosure problem, not a design problem. Stating the auto-renewal date, the cancellation steps, and the free trial's payment terms directly inside the checkout flow closes most of the gaps above. A terms and conditions generator built around these auto-renewal disclosure requirements keeps that language current without a manual rewrite every time a subscription flow changes.
Figure 6: Six years of app-specific dark pattern research and enforcement, largest studies and sweeps only. Sources: as cited throughout this post.
How do these numbers compare across studies?
| Study | Sample | Headline finding | Year |
|---|---|---|---|
| ICPEN Mobile and Online Games Sweep | 439 games | 60% send nagging reopen alerts | 2025 |
| Divergent Deceptions (iOS vs Android) | 143 apps, both platforms | iOS averages 10.3 instances per app vs 9.8 on Android | 2025 |
| ICPEN and GPEN Subscription Sweep | 642 traders | 75.7% use at least one dark pattern | 2024 |
| UIGuard dataset | 1,023 apps | Largest academic app-level dataset built to date | 2023 |
| University of Zurich crawl | 240 apps | 95% contain at least one dark pattern | 2019 to 2020 |
Source: as cited in each section above.
The Bottom Line
No single audit answers "how many apps use dark patterns" with one clean percentage, because the studies measure different things: some count games, some count subscription sellers across apps and websites together, and some count UI screens rather than whole apps. What every recent measurement agrees on is direction, not degree. 60% of games audited in 2025 nag players back in, most loot box mechanics go undisclosed, and three-quarters of subscription sellers checked in 2024 used at least one dark pattern. The oldest, cleanest app-level percentage still on the books, 95% of 240 apps, is now several years old, and the honest answer for a fresher all-category figure is that nobody has run that exact study again at the same scale. For a business shipping an app with in-app purchases, a free trial, or a subscription, the practical fix sits mostly in disclosure: state the terms plainly where the purchase happens, rather than relying on players not to notice.
Frequently Asked Questions
How many apps use dark patterns? There is no single definitive percentage covering every app category, but the largest recent audit, the International Consumer Protection and Enforcement Network's June 2025 sweep of 439 mobile and online games, found 60% sent nagging reopen notifications and 46% included loot box mechanics. The largest academic app-level percentage on record is 95% of 240 popular Android apps, from a 2019 to 2020 Google Play crawl that has not been repeated at a comparable scale since.
Do more dark patterns show up on iOS or Android apps? iOS, slightly. The same 143 mobile apps, reviewed on both platforms in a 2025 study published in Behaviour and Information Technology, carried 1,477 dark pattern instances on iOS versus 1,398 on Android, a statistically significant difference the researchers linked to app store guidelines and a rise in A/B-testing-driven design changes.
How common are loot boxes in mobile games? 46% of the 439 games examined in ICPEN's June 2025 sweep included loot box mechanics, randomized rewards players can buy with real or virtual currency, but only 14% disclosed that on their app store listing page, and 62% of the loot box games carried an age rating suitable for children.
Are dark patterns in apps illegal? It depends on the specific practice and jurisdiction, but regulators are treating them as enforceable violations. A companion 2024 ICPEN and GPEN sweep of 642 subscription sellers found 75.7% used at least one dark pattern, and consumer protection agencies including the FTC have brought cases over deceptive app and website design under existing law.
Where the Numbers Come From
- International Consumer Protection and Enforcement Network (ICPEN). (2025). "Mobile and Online Games Sweep Public Report," 30 June 2025. 439 games examined by 22 member agencies; 60% sent nagging reopen alerts, 46% included loot boxes, only 14% disclosed them.
- International Consumer Protection and Enforcement Network (ICPEN), with the Global Privacy Enforcement Network (GPEN). (2024). "Dark Patterns in Subscription Services Sweep Public Report," 2 July 2024. 642 traders examined; 75.7% used at least one dark pattern, 66.82% used multiple.
- Li, W., Flatla, D. R., and Arndt, F. (2025). "Divergent deceptions: comparative analysis of Deceptive Patterns in iOS and Android apps." Behaviour and Information Technology. Reviewed the same 143 apps on both platforms; 1,477 instances on iOS versus 1,398 on Android.
- Di Geronimo, L., Braz, L., Fregnan, E., Palomba, F., and Bacchelli, A. (2020). "UI Dark Patterns and Where to Find Them: A Study on Mobile Applications and User Perception." Proceedings of the 2020 CHI Conference. 240 Android apps crawled from the US Google Play Store in 2019; 95% contained at least one dark pattern.
- Chen, J. et al. (2023). "Unveiling the Tricks: Automated Detection of Dark Patterns in Mobile Applications" (UIGuard). Proceedings of the 36th ACM Symposium on User Interface Software and Technology (UIST 2023). Dataset spanning 1,023 mobile apps, 1,660 flagged dark-pattern instances.
- Chen, J. et al. (2026). "From Exploration to Revelation: App-Level Context-Aware Deceptive Pattern Detection for Mobile Applications" (AppRay). Preprint, arXiv:2411.18084. 100 apps across 22 categories, 2,185 flagged instances of 16 types.
Note: All figures verified as of July 2026. App-level dark pattern audits and datasets are refreshed as new sweeps and academic studies publish; expect this page to be updated at least twice a year as ICPEN, GPEN, and academic research teams release new figures.
