When AI Fights Scammers: Inside Google and Apate's New Defenses

Cybercriminals are scaling up their operations with artificial intelligence, using deepfakes and automated text scripts to squeeze victims out of hundreds of millions of dollars. In response, two distinct defense fronts are emerging: consumer-facing device protection and proactive corporate scam-baiting.

How Google and Apate are changing fraud detection

According to TechCrunch, Australian startup Apate takes an aggressive approach by deploying thousands of undercover AI agents that masquerade as vulnerable targets. These bots engage fraudsters across calls, emails, and phishing links, wasting 17 years of scammers' time over 2.5 million conversations while harvesting malicious wallet addresses and phone numbers for banks.

Meanwhile, as reported by The Verge, Google is tackling the issue directly on mobile devices. The upcoming 'enhanced protection' mode in Google Messages uses cloud-based Gemini models to evaluate hidden motives in conversations with unknown senders, processing user data in an isolated, encrypted environment.

How it works in practice

For everyday users, activating mobile safeguards requires opting into new device settings:

• Enable Enhanced Protection: In Google Messages, opt into the new cloud-based Gemini setting designed to analyze slow-burn conversation tactics from unrecognized numbers.
• Url Inspection: On-device Gemini Nano analyzes unfamiliar URLs in real time before links are opened to block malicious phishing sites.
• Call Forwarding Defense: Android devices now block unauthorized apps from enabling call forwarding and notify users instantly when attempts occur.

Veredito: who should use these tools?

These defense layers target two entirely different battlefields. Apate's platform is strictly for enterprise clients like banks and telecom operators seeking proactive threat intelligence. For individual professionals and remote workers, Google's on-device and cloud-based Gemini filters offer a frictionless shield against everyday mobile social engineering without requiring manual intervention.