Autoblog Forensics and Attribution
Aaron Koehl · 2018
Autoblogging is the process of running automated scripts against established content sources or syndicated feeds and compiling the information into blog sites. Unfortunately, unlike mashups which seek to enhance the Internet experience by coalescing disparate content sources in novel ways, a criminal ecosystem is evolving to profit off of the associated technologies. By scraping content en-masse from legitimate ad supported websites and removing the ads, attribution, and by rewriting links, a malicious site can impersonate a legitimate site, yet with their own ads for profit. We introduce a forensic approach to help solve the problem of traceability when content is stolen en masse from one site and hosted on another. Content is first minified and then watermarked repeatedly with unique identifiers using a sparse encoding tailored to the user's available steganographic bandwidth. When scraped content is discovered, the encoded identifiers can be traced back to specific events in server logs, as a powerful aid in identifying perpetrators and to help combat this practice.