Apple is using some pretty eye-popping numbers to make a simple point about 2025: the App Store is trying very hard to be a safer place for both users and developers, and it is leaning heavily on AI plus human review to get there. Behind the marketing gloss, there is a clear story of scale, constant cat-and-mouse with fraudsters, and increasingly aggressive moderation.

In 2025 alone, Apple says it blocked more than $2.2 billion in potentially fraudulent transactions on the App Store, bringing the total to over $11.2 billion stopped in the last six years. That is not money Apple lost; it is money the company claims never reached scammers because its systems flagged suspicious activity before charges went through. At the same time, Apple’s Trust and Safety teams were playing defense on every front: fake accounts, dodgy developers, pirate app stores, shady in-app payments, and even manipulated ratings and reviews.

A big chunk of that work starts before a single app ever hits a user’s device. Apple’s App Review team handled more than 9.1 million submissions in 2025, from brand-new apps to routine updates. Out of that firehose, over 2 million submissions were rejected because they failed to meet Apple’s App Store Review Guidelines on things like privacy, security, spam, or misleading behavior. That includes more than 1.2 million new apps and nearly 800,000 app updates that never made it through the front door.

As AI tools make it easier to churn out new apps and code, Apple has quietly scaled its own AI to keep pace. App Review has long used a mix of human judgment and machine learning, but Apple now emphasizes a “multilayered” approach that spots suspicious patterns at scale, flags risky updates, and highlights cases where human reviewers should take a closer look. This is not just about catching obvious malware; Apple is watching for apps that try to play nice at first and then flip into scam mode later. In 2025, nearly 59,000 apps were removed after launch for “bait-and-switch” behavior, such as a harmless-looking calculator that later morphs into a fraudulent financial app.

The company is also blunt about one growing headache: account fraud. Fake accounts are gold for scammers, letting them spam users, game rankings, or pump up fake reviews at scale. In 2025, Apple says its systems rejected more than 1.1 billion fraudulent customer account creation attempts outright, often driven by bot networks designed to flood the platform. On top of that, it deactivated 40.4 million existing customer accounts for fraud or abuse, suggesting that bad actors are constantly probing the defenses instead of just bouncing off at sign-up.

Developers are under scrutiny too. In 2025, Apple terminated around 193,000 developer accounts over fraud-related concerns, and blocked more than 138,000 additional developer enrollments it considered suspicious. The message is fairly clear: if you are trying to sneak a scammy app into the ecosystem, Apple wants to cut you off at the account level, not just play whack-a-mole with each individual app. That push extends beyond the App Store itself into the murky world of pirate storefronts, where cloned and modified apps are often loaded with malware or used to funnel users into gambling and pornography services. Apple reports detecting and blocking about 28,000 illegitimate apps on these off-platform storefronts last year.

Even outside the official store, Apple is trying to keep a leash on what gets installed on iPhones and iPads. Over the last month alone, the company says it prevented 2.9 million attempts to install or launch apps distributed illicitly outside the App Store or approved alternative marketplaces. That figure matters in the context of regulatory pressure, especially in regions where Apple is being pushed to support sideloading or additional app stores: Apple is quietly signaling that the wider you open the door, the more aggressively you need to police what walks through it.

Ratings and reviews, which most people casually scroll through before downloading, are another battleground. In 2025, Apple processed more than 1.3 billion ratings and reviews on the App Store. Using a combination of AI and human moderators, it says it blocked almost 195 million fraudulent or inauthentic ratings and reviews from ever appearing. That covers classic tricks like paid five-star reviews, review farms, and coordinated campaigns to tank a competitor’s score. For users, inflated ratings can make a scam app look legitimate, and for developers, fake negative reviews can bury an honest app. Apple also blocked roughly 7,800 deceptive apps from showing up in App Store search results and another 11,500 from appearing on top charts, to prevent discovery from being gamed behind the scenes.

When it comes to actual money changing hands, Apple leans heavily on its payment stack: Apple Pay and StoreKit. The company says more than 680,000 apps now rely on its in-house payment technologies to sell goods and services. These payments are wrapped in industry-standard encryption, but the more interesting piece is how Apple uses machine learning to watch the flow of transactions in real time. In 2025, that setup helped block over $2.2 billion in fraudulent transactions, stop more than 5.4 million stolen credit cards from being used in the App Store, and ban nearly 2 million user accounts from making future transactions. For users, that can mean cards being declined “for your protection” when something looks off; for Apple, it is a way to show banks and regulators that its closed system has real security benefits.

The safety story is not just about stopping professional scammers. Apple also likes to highlight the controls it gives families, especially around kids. The dedicated Kids category on the App Store comes with stricter rules for age ratings and in-app advertising, and Apple says it rejected more than 5,000 apps from appearing in that category in 2025 because they did not meet those standards. Developers can tap features like the Declared Age Range API and PermissionKit to build experiences that adjust content and permissions based on age, which is increasingly important in markets where children’s online privacy is under sharper regulatory focus.

On the user side, Apple pitches tools like Screen Time and Ask to Buy as simple ways to stay in control. Parents can set app limits, define downtime, and put content and privacy restrictions in place, while Ask to Buy lets them approve each new app download or in-app purchase on a child’s device. And if something still slips through, Apple directs users to its “Report a Problem” channel, where suspect apps or transactions can be flagged for investigation and potential removal. It is a reminder that even with all the AI and automation, Apple still relies on user reports to catch the edge cases algorithms miss.

All of these numbers also serve a bigger narrative purpose. Apple has been under pressure from regulators, developers, and rivals over how tightly it controls the App Store and how much it charges for that privilege. By publishing detailed fraud and safety stats, and highlighting growth from previous years, Apple is essentially arguing that its control is not just about revenue-sharing, but about maintaining a secure, curated environment at enormous scale. Independent analyses have noted that Apple has become more aggressive about delisting apps for fraud or policy violations over the last few years, even as critics point out that deceptive apps still occasionally slip through.

From a user’s perspective, the takeaway is fairly straightforward: the App Store is not perfect, but there is a lot happening behind the scenes that you never see. Every time you tap “Get” or confirm a subscription, there are machine learning models and review teams trying to make sure that download is not malware, that rating is not bought, and that transaction is not happening on a stolen card. For developers, especially smaller ones, the same systems that sometimes feel like a slow-moving gatekeeper are also the barrier that keeps scammers, copycats, and pirate storefronts from turning the marketplace into a free-for-all.