Noise contamination

Modifying the content of spam with the aim of bypassing mail system filters.

Most email service providers automatically scan the text of incoming mail to weed out scam or unsolicited messages. To circumvent security barriers, cybercriminals add “noisy” words or symbols to the message text to mislead machine-learning algorithms, while keeping the content of the message readable to the recipient.

Examples of noise contamination:

• Using Latin characters that look the same in Cyrillic

• Using numbers instead of letters and vice versa (for example, 0 instead of O).

• Separating letters in a word with spaces or dashes

• Adding unreadable or invisible characters to the message

• Adding “clean” links, extracts from literary works, or other non-suspicious text to the message