Policy and regulation illustration
Chrome Security, working with Firebase Cloud Messaging and Safe Browsing, published a technical account on 11 August 2026 of how it is trying to stop websites from flooding people with deceptive push notifications. Hannah Buonomo, Jonathan Li and Nidhi Davawala wrote that abusive notifications have been used to push malware, harvest personal information and solicit fraudulent payments. Google said that after a multi-year effort, unwanted notifications on Android fell by more than 7 billion a day in the first quarter alone. That volume reduction is Google’s measurement. The team described a defence-in-depth model in which overlapping controls cover the whole notification lifecycle. Chrome automatically revokes notification permission from sites a user has not engaged with recently, and from sites that have repeatedly triggered suspicious-notification warnings. Users can still restore permissions in Safety Hub. Google said it also built behavioural detection aimed at networks of sites that coordinate abusive notifications, including signals from service-worker activity, so that permissions can be removed even when the site content does not look obviously malicious. On the server side, Firebase Cloud Messaging now applies rate limits to the Push API. Google said it evaluates sites using message volume relative to time spent on the site, how often they prompt for permission, and general engagement. Domains judged disruptive are limited to 1,000 messages per minute and receive HTTP 429 responses if they exceed that threshold. The company said the limits scale with repeat offences and reset only after a period of non-disruptive behaviour, raising the cost of large-scale abuse. Chrome has also changed how permission is granted. Google said it recently updated the notification grant experience on Android phones so people can decide whether a site may send notifications without being pulled out of their browsing. That builds on one-tap unsubscribe on Android, which was meant to make it easier to drop a noisy site. Google argued that the combined work also cuts unnecessary background activity and battery use, though those side effects are presented as the company’s assessment rather than a third-party study. Abusive notifications remain a social-engineering problem as much as a browser one. A warning that looks like a bank, delivery tracker or antivirus alert can still succeed if the user grants permission on a lookalike site. Chrome’s automatic revocations help after the fact; they do not replace caution at the moment a prompt appears. What to do • On a computer, open chrome://settings/content/notifications and remove sites you do not remember allowing. On Android, use Chrome’s menu, Settings, then Notifications. • Decline notification prompts from sites you visited once, especially after a pop-up or “you are a winner” page. • If a notification asks you to call a number, install software or pay a fee, ignore it and open the real service by typing the address yourself. • Small organisations should not prompt every visitor for notifications, and should review which marketing tools register a service worker. • Keep Chrome set to update automatically so permission and Safe Browsing changes reach the device. Primary source: Google Chrome Security blog, “The multi-layered defenses that harden Chrome against abusive notifications”, 11 August 2026.

Sources

  1. Google / Chrome Security — The multi-layered defenses that harden Chrome against abusive notifications2026-09-24

Cyber71 links the primary material behind each report so readers can check it directly. External pages are not under our control and may change.

All news

Need help with a security issue?

Describe the situation and we will open a private case for follow-up.

Get Security Help