A Big‑Numbers Milestone in App Store Fraud Prevention
Apple’s latest figures underline just how large the App Store fraud problem has become—and how aggressively it is being tackled. In 2025 alone, Apple says it blocked over USD 2.2 billion (approx. RM10.1 billion) in potentially fraudulent App Store transactions, bringing the six‑year total to more than USD 11.2 billion (approx. RM51.7 billion). Those interventions are paired with tighter control over what reaches users in the first place: the App Review team evaluated more than 9.1 million submissions and rejected over 2 million apps and updates that failed to meet App Store Review Guidelines. These rejections spanned issues from privacy violations to deceptive behavior. Apple positions these numbers not just as a defense of consumers, but as protection for legitimate developers who depend on a trusted marketplace that draws over 850 million weekly visitors across 175 storefronts.
How Machine Learning Supercharges Human App Review
Behind those statistics sits a hybrid review model that blends automation with human judgment. As AI‑assisted development tools flood the App Store with more submissions, Apple has scaled its systems by weaving machine learning into every stage of App Review. Algorithms comb new apps and updates for complex malicious patterns, analyze similarity with known bad software, and flag unusual behavior changes between versions. These automated checks accelerate triage so that human reviewers can spend time on the riskiest or most ambiguous cases instead of routine screening. The approach is particularly crucial for catching sophisticated fraud, such as apps that behave like harmless games or utilities during review but later morph into tools for financial scams. By spotting suspicious post‑review changes quickly, Apple can remove problematic software before it causes widespread harm.
Shutting Down Fraudulent Accounts at Scale
App Store fraud prevention increasingly starts long before an app appears on a storefront. Apple’s Trust and Safety teams are using AI and analytics to shut down fraudulent account operations at industrial scale. In 2025, systems rejected 1.1 billion fraudulent customer account creation attempts, blocking bad actors at the point of entry. A further 40.4 million customer accounts were deactivated for fraud or abuse. On the developer side, Apple terminated approximately 193,000 developer accounts over fraud concerns and rejected 138,000 attempted developer enrollments. These measures target bot networks and fake identities used to spam users, manipulate charts, flood ratings with fake reviews, or repeatedly submit repackaged malicious apps. By cutting off these pipelines early, Apple reduces the volume of dangerous software that ever reaches the formal App Review queue, easing pressure on reviewers and strengthening overall platform integrity.
Detecting Malicious Apps, Clones, and Bait‑and‑Switch Tactics
Even after rigorous screening, some malicious developers attempt to sneak harmful behavior into the App Store. Apple’s data shows a tightening grip on these tactics. In 2025, nearly 59,000 apps were removed after engaging in bait‑and‑switch maneuvers—launching as benign puzzles or calculators, then modifying their code post‑approval to facilitate financial fraud or other abuse. App Review also rejected over 22,000 submissions that hid undocumented features, more than 371,000 that copied other apps or misled users, and over 443,000 for privacy violations. Outside the official storefront, Apple detected and blocked 28,000 illegitimate apps on pirate platforms, ranging from malware and gambling apps to pornography and pirated versions of legitimate titles. The company also reports blocking 2.9 million attempts in a single month to install apps from illicit channels, underscoring how fraudulent app detection now spans both official and unofficial ecosystems.
Why the Dual Approach Matters for Users and Developers
The combined use of AI and human review is changing both the scale and precision of App Store security measures. Machine learning enables real‑time fraud detection across billions of signals—account sign‑ups, code patterns, review behavior—while experts interpret edge cases, refine policies, and spot novel threats. This dual approach has produced record levels of app submission rejection for harmful or low‑quality software, while still admitting over 306,000 new developers to the platform in 2025. Apple also extends checks to TestFlight, blocking more than 2.5 million prerelease builds over security or fraud concerns before they reach broader testers. For users, the result is a more trustworthy storefront with fewer malicious apps and manipulated ratings. For developers, it means a marketplace where genuine products are less likely to be drowned out by clones, scams, and pirated copies.