Gmail was better to block phishing attacks through machine learning

After a recent phishing attack affecting about one million Gmail users, the service increased the level of security of the mail. According to the developers, the new machine learning model that is able to recognize spam and phishing with an accuracy of 99.9%. The result looks even more impressive when you consider that this “garbage” accounts for up to 50-70% of the incoming emails.

After a recent phishing attack affecting about one million Gmail users, the service increased the level of security of the mail. According to the developers, the new machine learning model that is able to recognize spam and phishing with an accuracy of 99.9%. The result looks even more impressive when you consider that this “garbage” accounts for up to 50-70% of the incoming emails. Google also introduced a technology for the early detection of threats, which is already available to users of the corporate Suite package G.

Protected

Gmail algorithms now take into account the identity of the sender, and use heuristic analysis on the basis of characteristics that may indicate spam or malicious message content. This allows you to identify even previously unknown threats. The service may delay delivery of suspicious emails and to conduct a detailed inspection. According to the company, therefore, is taken not more than 0.05% of the incoming messages.

To prevent information leak from the company G Suite users will also see a warning if the source is not included in the list of contacts, and his mailing address was on a different domain. Among other security features, a Gmail — display warnings in the browser when navigating to potentially dangerous websites (implemented on the basis of Safe Browsing) and question marks instead of avatars of the user in the emails from suspicious or untrusted sources.

Source: Google