Child Exploitation (CE) imagery continues to be distributed publicly on the Internet. More than three-quarters of child sexual abuse imagery identified by the Internet Watch Foundation in 2014 was located on common domains, and not hidden. The accessibility of illicit media online is a significant concern for both parents and governments. Despite this, little is known about public CE websites.
Westlake, Bouchard and Girodat studied how websites providing CE material operate online. They attempted to determine how open CE website operators were about their illegal activities and whether they used any tactics to hide the content. They analyzed 634 websites distributing child sexual exploitation material, or hyperlinking to such a website. These websites contained at least one known CE image, or seven known CE keywords for a hyperlinked webpage. They then compared their review to an automated study of the same websites.
The overall purpose of this study was to determine how obvious it was that a website was explicitly CE-focused. The answer to this question lies in how well the websites were able to conceal their purpose from an automated data collection process. Consequently, there were two data collections for this study. The first was by an automated webpage collection tool. The second was a manual investigation of each website’s homepage conducted over a period of two months, with a 14-month follow-up.
They found that many of the explicitly CE-focused websites identified manually were also identified automatically, suggesting that CE-related websites do little to hide their purpose. The presence of CE images or an explicit CE focus did not impact the survival of a website. 14 months after the initial study, 80% of the CE websites were still online, compared with 84.9% of all websites, including those not related to child exploitation.
The websites with CE images did not try to hide their intentions. Despite this, they were no more likely to fail than other websites. This reinforces the assumption that conducting illicit activities in public on the Internet generally does not increase the risk of failure.
For autonomous data collection tools to be effective in detecting CE websites, they need to be provided with proper guidance. Language and terminology evolve over time. The use of ‘code words’ is perhaps still even more fluid. Offenders must adapt quickly to avoid terms that have become commonplace. This means that autonomous tools need to extend beyond searching for only the code words used specifically to identify CE content to more descriptive language.
Child exploitation is no longer a hidden realm only accessed by sophisticated, technological masters. CE websites are easy to detect. This highlights the jurisdictional, privacy, and identification issues for law enforcement. Through understanding the layout, accessibility and the readily available content of CE websites, governmental agencies can more effectively use current tools to remove the content from the Internet.
Child Exploitation Websites don’t appear to need to hide. Law enforcement seem to need better tools to overcome the current jurisdictional, privacy, and identification barriers.