According to TechCrunch , Google said its machine learning technology is now able to detect spam and phishing messages with 99.9% accuracy.
This means that only one in every thousand messages inadvertently reaches the Gmail inbox. Currently 50-70% of the messages the Google e-mail service receives are spam.
For messages reaching the user, the company uses its Google Safe Browsing tool to detect links to malicious URLs – alerting the user at the time of the click.
With respect to phishing detection, a new system deployed on the platform will delay the receipt of suspicious messages for deeper analysis.
Google said only about 0.05% of messages should be affected by the delay of this new verification.