Seven billion undesirable notifications a day is plenty of noise to chop from Android units.
Google says Chrome decreased undesirable notifications by greater than 7 billion per day within the first quarter of 2026 by combining permission revocation, abuse detection, sender price limits, and on-device machine studying.
The corporate is now utilizing a number of defenses collectively. Chrome can revoke notification permissions, detect coordinated habits throughout web sites, flag suspicious notification content material, and restrict high-volume senders via Firebase Cloud Messaging (FCM).
That method issues as a result of it shifts a few of the burden away from customers, who would in any other case have to acknowledge each suspicious notification themselves.
A number of approaches to deal with a single downside
Filtering out such a quantity of unwanted notifications is a big job, and Google has now revealed the behind-the-scenes course of for doing so.
According to the company’s blog, Chrome is using a “Swiss cheese” mannequin to cope with undesirable notifications.
Google’s first line of protection is the infrastructure behind the notifications. Chrome can mechanically revoke notification permissions from websites that seem abusive, repeatedly set off warnings, or have change into inactive, slicing off their skill to maintain pushing alerts to a person’s machine.
Google can be on the lookout for indicators that a number of web sites could also be working collectively.
Google can use indicators akin to service-worker exercise to establish coordinated habits, whereas its FCM provides one other management by rate-limiting websites that ship notifications at extreme volumes. Web sites could be restricted to 1,000 messages per minute, with repeat offenders dealing with tighter restrictions.
The corporate can be making it simpler for customers to show off notifications with out digging deep inside Chrome. Push notifications will now embody an choice to show them off instantly.
Extra Google protection
Chrome is constructing upon earlier work
The 7-billion determine displays greater than Google’s newest modifications.
Chrome’s notification defenses have been developing for months, together with on-device Machine Studying that may analyze notification content material domestically somewhat than sending it to Google, an method that helps hold the notification content material end-to-end secured.
However Android’s {hardware} fragmentation leaves one query unanswered: how broadly can these native protections really run?
Google has not specified whether or not the on-device ML layer reaches funds and older units as broadly as newer {hardware}, which might have an effect on how constantly customers obtain the identical degree of safety.
The corporate retains refining how safe Chrome notifications stay by now including its latest multi-layered method. In its personal phrases, the “aim is to make sure that if abuse slips via one layer, one other is there to catch it.”
The person nonetheless has a task to play
Even with all these defenses, Google can’t make the notification channel risk-free.
A notification can nonetheless be misleading with out triggering each automated protection. Meaning the person’s judgment stays the ultimate layer of safety, making person consciousness as necessary as Google’s protections.
There may be additionally the chance that automated security systems could get it wrong. Google hasn’t disclosed an incident through which these notification defenses made a particular mistake. However any system that mechanically identifies and restricts content material or web sites has to cope with false positives. A authentic web site might be caught by an excessively aggressive protection, irritating customers who really need its notifications.
Different Information: Google says Gemini has surpassed 1 billion monthly active users, marking a significant milestone because it competes for scale within the quickly rising client AI market.
