Android Boosts AI Defenses Against Call & Text Scams

Millions of Scams Caught Each Month

Google says its AI already flags around 5 million suspected scam calls and texts every month inside the Phone and Messages apps, a scale that comes partly from recent additions like Messages’ edit history feature. The new round of protections builds directly on that existing detection layer rather than replacing it.

Messages Gets a Deeper Look at Suspicious Texts

An optional enhanced protection mode is coming to Google Messages that lets its AI examine message content more closely, helping it spot a scammer’s intent faster when a message comes from an unknown sender. Google says this analysis happens with encryption on its servers rather than locally, and the mode will also screen for malicious links and unrecognized numbers.

The rollout is staggered:

• Enhanced Messages protection arrives on Pixel within the next few weeks
• Chat notification scam detection extends later this year to Motorola, OnePlus, and Xiaomi devices

Calls and Screen Sharing Get Locked Down Too

Android is also adding proactive call-forwarding protection that stops apps from silently rerouting calls, a tactic scammers have used to intercept one-time passcodes and take over accounts. Anyone attempting to set up call forwarding will now get an explicit notification first. It’s a narrower fix than the broader call-understanding work Google has been testing with Gemini on calls, but it closes a specific and well-documented hijacking method.

Screen sharing picks up a privacy fix as well. If someone is sharing their screen and needs to type a password or other sensitive field, Android will automatically hide it so it never shows up on the shared display.

Why Now

Google commissioned a YouGov survey of 4,500 adults and found 40 percent were very or extremely concerned about mobile fraud, numbers the company is using to justify the rollout’s timing. It also lands right as former Samsung Messages users finish migrating to Google Messages following that app’s shutdown in July, putting a fresh batch of users inside the ecosystem these protections are meant to cover. Google laid out the full rollout in a blog post published Wednesday.

The Privacy Tradeoff Behind “Enhanced” Protection

Google’s emphasis on server-side encryption is doing a lot of work to make AI content scanning sound harmless, but the plain fact is that opting into enhanced protection means Google’s systems are reading the substance of your texts, not just metadata like sender and link patterns. That may well be a reasonable trade for people getting hammered by scam texts, but it is a meaningfully bigger scope of access than the keyword and number-based filtering most users assume “spam protection” means, and Google’s announcement does little to spell that distinction out.