Google Play used AI to assist block 1.75 million dangerous apps in 2025

Google has announced that with the assistance of AI, it blocked 1.75 million apps that violated its insurance policies in 2025, considerably down from 2.36 million in 2024. The decrease numbers this 12 months, it mentioned, are as a result of its “AI-powered, multi-layer protections” are deterring dangerous actors from even making an attempt to publish dangerous apps.

Google mentioned it now runs greater than 10,000 security checks on each app and continues to recheck them after they’re printed. Its use of the newest generative AI fashions helps human reviewers uncover malicious patterns extra rapidly, it added. The corporate additionally blocked 160 million spam rankings, stopping a median 0.5-star score drop for apps focused by evaluation bombing. Lastly, Google stopped 255,000 apps from gaining extreme entry to delicate person knowledge in 2025, down from 1.3 million the 12 months earlier than.

In the meantime, Google Play Defend, the corporate’s Android protection system, sniffed out over 27 million new malicious apps, both warning customers or stopping them from operating. The corporate added that Play Defend’s enhanced fraud safety now covers 2.8 billion Android units in 185 markets and blocked 266 million dangerous “side-loading” set up makes an attempt.

“Initiatives like developer verification, obligatory pre-review checks, and testing necessities have raised the bar for the Google Play ecosystem, considerably decreasing the paths for dangerous actors to enter,” the corporate mentioned its weblog. “This 12 months, we’ll proceed to put money into AI-driven defenses to remain forward of rising threats and equip Android builders with the instruments they should construct apps safely.”

Google has steadfastly justified its comparatively excessive charges on app purchases and subscriptions by touting its investments in app security. Nonetheless, its Play retailer has been below strain from regulators in Europe and different areas that declare it quantities to a monopoly. Final 12 months, the corporate modified its charge construction for builders utilizing different fee channels, however EU regulators just lately claimed the corporate still isn’t complying with Digital Markets Act laws.