E-mail attacks on the rise

MArchive full series mail archiving servers and MDS/MDispersion mid-to-high-end series mail servers adopt AI automated spam protection reasons, according to research statistics conducted by the US Cybersecurity and Infrastructure Security Agency in December 2022, at least 8 out of every 10 companies Every now and then, employees of businesses fall victim to phishing by hackers.
Relevant data shows that most Internet users are unprepared for information security and emails. However, today's hackers attack emails in a variety of ways, such as: APT and BEC and other highly targeted attacks, which make administrators and general users unable to defend effectively.
When hackers use various methods to attack through the information network, when hackers carry out "24 hours a day Zero day vulnerability", in the virus scanning websites such as International Virus Total and various Anti-virus software, fake phishing websites or some attached files are often regarded as harmless letters. Because the zero-day attacked websites and some attached files still belong to relevant databases and have not recorded new types of harmful letters, administrators and users Or the computer's built-in or anti-virus software fails to detect some high-potential harmful websites and virus files, which will still lead to serious consequences later.
HERHSIANG Information has introduced the AI Spam automated spam filtering mechanism, which can effectively isolate or block harmful emails after comparing and learning from the database, thereby preventing harmful emails from entering the user's mailbox, which can effectively protect the security of the user's related equipment Safety.

Features:
AI Spam function introduction
Collect a huge Chinese word segmentation database. The calculation basis of AI Spam is traditional/simplified Chinese word segmentation database. The difference between Chinese and English is the key to the accuracy of word segmentation to determine the email intent analysis. For example: Go to Global Information on the internet
Segmentation determines the semantic classification of the above-mentioned words. The R&D unit has done research and shows that after entering AI SPAM calculations between continuous words and segmented words, the accuracy difference is 1~2%. The slight difference is reflected in the accuracy of the final email judgment To a certain extent, with the help of the Chinese word segmentation database established by the research and development unit, AI Spam is more accurate and simple in the calculation of mail classification.

AI Word Analysis module
The AI Spam imported by HERHSIANG Information disassembles the subject and content of the email and then enters the comparison action. The relevant Chinese word segmentation database has fully complied with the usage habits and usage of traditional and simplified Chinese. Enter the AI algorithm program classification comparison analysis after the words. AI Spam database statistics have built more than 1 million classified emails. The database can be divided into five categories: normal/notification/e-mail/junk/phishing emails, After AI automatic comparison and analysis, new emails will be automatically classified into five categories according to relevant characteristics. With the help of email vocabulary and semantic classification, users can be asked to "immediately" click on the link URL in the email or open the "full file" 24 hours a day Zero day vulnerability" mails are effectively blocked outside the relevant equipment to prevent them from entering and endangering the user's equipment.

Content filtering analysis module
In addition to classifying harmful emails, the AI Spam automatic spam protection system will check whether the URL links and attached files in the inner text of the letter contain known viruses and harmful URLs in the system database.
In addition, through the HERHSIANG Sandstorm database, known phishing/trojan horse emails are blocked. There are more than 2 million phishing URLs and 1 million Trojan horse data files in Sandstorm. For spam hazard emails that cannot be identified by the content, for example, For short content emails/graphic text emails, etc., use BAYES (Bay's Theorem) rules to assist in judgment.

Analysis of letter sending behavior
Hexiang Information has introduced a variety of module judgment methods to reduce the probability of misjudgment. After AISPAM/Sandstorm and BAYES (Bayes' theorem) rules and other first-level classifications based on email content, the AI SPAM system is developed by the research and development unit Analyze and judge the behavior of the sender, find out the large amount of forwarding/illegal forwarding sent by the robot computer, illegal forwarding through free mailboxes or even fake domains and other unreasonable sender behaviors, and at the same time use the basic verification mechanism SPF Sender domain lookups such as /DKIM/DMARC are also processed.

These emails sent through abnormal channels have a characteristic. They will definitely not be normal emails, and will be classified as e-mails and spam by the system. When the user reconfirms the sender and email from the notification letter After the subject is safe, you can retrieve the safe letter from the AI SPAM automated spam protection system.

Analysis of letter sending behavior module
HERHSIANG Information has introduced a variety of module judgment methods to reduce the probability of misjudgment. After AI Spam/Sandstorm and BAYES (Bayes' theorem) rules and other first-level classifications based on email content, the AI SPAM system is developed by the research and development unit Analyze and judge the behavior of the sender, find out the large amount of forwarding/illegal forwarding sent by the robot computer, illegal forwarding through free mailboxes or even fake domains and other unreasonable sender behaviors, and at the same time use the basic verification mechanism SPF Sender domain lookups such as /DKIM/DMARC are also processed.

These emails sent through abnormal channels have a characteristic. They will definitely not be normal emails, and will be classified as e-mails and spam by the system. When the user reconfirms the sender and email from the notification letter After the subject is safe, you can retrieve the safe letter from the AI Spam automated spam protection system.

AI Spam Automated Spam Protection keeps improving

HERHSIANG Information has introduced a new AI Spam function module, which can make the judgment accuracy of mails as high as 99.5%. Through the judgment and analysis of mail content and sending behavior of the system function, it can automatically distinguish between normal and malicious or spam letters, and improve the level of spam letters It can also reduce the chance of users opening phishing letters.
In addition, hacker attack methods are also constantly improving, they will forge emails to be the same as real emails, so the MArchive mail archiving servers and MDS/MDisperison email host core system must constantly learn new attacks by itself or can be phished mail.
The research and development unit has included these attack and phishing features in the calculation logic database as a basis for judgment, so that the AI Spam system can use the continuous self-learning mechanism to face the phishing letters and spam attacks that are attacked 24 hours a day throughout the year. It can also quickly make accurate judgments, thereby effectively reducing spam and phishing letters to users, causing unnecessary losses.