International Journal on Criminology Volume 7, Number 2, Spring 2020 | Page 102
The Memex Project
Human Trafficking Network Investigations
The pioneer of the internet, Defense Advanced Projects Research Agency
(DARPA), part of the Department of Defense, launched Memex in 2014,
a major analytical program focused on developing methods for indexing
parts of the internet that were previously unsearchable by exploring trafficking
for sexual exploitation online. This multi-year program was motivated by both
a national security interest and a desire to develop search engine capabilities to
help law enforcement use information on the deep and dark web; information
that cannot be found by mainstream, surface web search engines. This program,
employing many of the country’s top computer scientists, cost nearly $67 million 17
and developed key tools to scrape and locate data on the deep and dark web.
Much of the data found by the DARPA research team on commercial sex
advertisements was located on the surface and deep web, as opposed to the dark
web, given the desire of commercial sex advertisers to reach the largest number of
customers. Customers could easily access advertisements for sexual services on
the surface web featured on the specialized sections of Craiglist and Backpage that
featured links to more exclusively sexual websites, such as Escortphonelist, escort
review sites existing behind paywalls on the deep web, such as TheEroticReview,
and over one hundred other identified websites containing erotic, massage, and
sexual advertisements. Advertisers of sexual and erotic services generally avoided
use of the dark web for several reasons. First, the dark web is not easily accessible
by the typical sex buyer. In order to access the dark web, the user must know the
specific web location of the site or marketplace they want to enter. Secondly, the
user must use a specialized web browser, such as Tor, to access the dark web, which
surpasses the computer skills of many possible customers. In general, use of the
dark web is in opposition to the desires of those seeking sexual services, who want
variety, choice, and the ease and speed of accessibility; therefore, advertisers prefer
to use sites on the surface or deep web.
A central objective of the Memex project was to harness open source intelligence
(OSINT) from escort advertisement websites and sex buyer review forums
by storing the intelligence in a searchable database and equipping investigators
with otherwise inaccessible information. This intelligence could more readily inform
investigators’ use of criminal intelligence contained in their sensitive, closed
databases. In order to accomplish this goal, Memex had to provide an efficient
means for searching large and diverse types of content that could not be analyzed
manually. Additionally, the data collected was frequently changed or manipulated
by the poster, in order to avoid detection. The collected and searchable OSINT
provided additional opportunities to cross-reference information generated by
17 C. Pellerin, “DARPA Program Helps to Fight Human Trafficking,” DoD News, Defense Media Activity,
January 4, 2017, https://dod.defense.gov/News/Article/Article/1041509/darpa-program-helpsto-fight-human-trafficking/.
93