According to Google statistics, Gmail inbox 50% -70% of the information is spam, and a considerable part of them are carefully planned to confuse the user. In this regard, Gmail hopes to learn through the machine technology to help users avoid the human factors caused by fishing phenomenon. When the AI detection system finds a suspicious message into the inbox, it will take a hysteresis on the message and then analyze it and use it to determine whether the destination message is a phishing message.
Gmail’s Internet abuse response technology product manager Andy Wen said that the new model they used was faster to adapt than the artificial system and could get better over time. Google said that lagging mail accounted for less than the total number of 0.05%, but taking into account the huge Gmail mail base, so even a small proportion of the impact of the user is still very large.
It is understood that Gmail automatic malware search system will be used from Google Safe Browsing machine learning technology, it will find suspicious fishing mail when the user issued a fishing warning and contact with the network attack related information link. Thanks to hundreds of spam messages sent to Gmail mailboxes, Google has built a set of artificial intelligence systems that can quickly identify malware and extort software information and automatically block it.
In addition, Gmail has also launched a feature for enterprise users, that is, when sending a message containing sensitive information issued a warning. But companies can choose to turn off this feature and early fishing detection.