Spammers started using artificial intelligence

Security experts believe that graphical spam is the first use of Artificial Intelligence technology by spammers, but this is only the tip of the iceberg.



Analysts from Forrester Research have published a detailed report on this issue. They believe that with the advent of new models that use AI, it will be impossible to fight with graphic spam with modern methods, because spamming neural networks are trained on the go. They generate millions of emails that are different from each other, some of which - and, therefore, millions of copies - in any case break through the filter.



Here is the same situation as in the CAPTCHA tests. The computer (in our case, the anti-spam filter) cannot read the text graphically, but the person (the recipient of spam) can. And graphic spam is not the only challenge we face.



Spammers have already begun to use various graphics variations, strikethrough and noisy images, audio and video files. If we start blocking each of these types of spam at the filter level, then we will experience failure. According to experts, you need to look at the root of the problem and deal with the fundamental properties of spam, such as links to third-party sites, because they are contained in almost every spam message.

')

One of the potentially successful ways to combat spam is social filtering, tagging spam in real time by the efforts of millions of users.



via Infoworld