Promise and Peril in Tech Approaches To Stopping Child Sex Trafficking

From surveying to surveilling to spreading information, one organization is showing promise and raising concern.


Children and teens are trafficked for sex and exploited for pornography, and the web amplifies this trade. The National Center for Missing and Exploited Children received 500,000 potentially explicit images of minors online in 2004. By 2011 it surged to over 17 million. An estimated 100,000 minors are "at risk" for trafficking.

Legislators try every year to address trafficking. Seventy-six bills and resolutions on human trafficking have been introduced in the 113th Congress so far. But government action regularly falls short of addressing underlying issues and has unintended consequences: Innocent people end up on lifelong registries, guilty people end up indefinitely detained in ineffective reform institutes, and vulnerable sex workers end up violated, displaced, and homeless at the hands of the police.

A Silicon Valley-based organization called Thorn is trying a more tailored, tech-centered approach. The group is low-key, but that shouldn't be mistaken for low-impact: Thorn is financed by celebrities like Demi Moore and has partnerships with major web companies like Google and Facebook. The group is implementing three strategies–surveying, surveilling, and spreading information–to reach out to kids who are susceptible to exploitation. To be clear, Thorn recognizes adults' rights to voluntarily engage in sex work, so that's outside the scope of their advocacy and action. Thorn gives some hope for the future of protecting minors until they're old enough to make these decisions on their own, but the organization's work also raises necessary questions about its own ethics and effectiveness.

The first "big picture" survey may miss key details

Thorn Director of Programs Claire Schmidt said at Stanford University's Liberation Technology series in January, "There aren't that many people in the traditional non-profit sex trafficking prevention world that are actually equipped with the tools and technical know-how … to even understand exactly what's happening in these environments." So her organization conducted a national survey, which is significant, as it's one of the first pieces of research at scale on the trade. It gives them a clearer sense of potential solutions for previously unknown or misunderstood issues such as traffickee-trafficker dynamics, especially since the Internet has quickly and fundamentally changed commerce and communication. For example, they found that "70 percent [of trafficked children are] sold online at some point in their trafficking situation," but that "83 percent of respondents met their trafficker in person" which "did run contrary to a lot of anecdotes [Thorn] was hearing." This has helped shape the initiatives discussed below.

Statistics have their limits, though. For one, the study is skewed by a lack of input from males, who are less likely to discuss trafficking experiences. And data points don't necessarily create an accurate image of reality. Maggie McNeill, a sex work writer and advocate and Reason contributor says that people labeled "trafficker" or "pimp" rarely fulfill cane-carrying stereotypes. She explains that cops will arrest a sex worker's boyfriends and charge him for trafficking. Or, when police find a group of prostitutes living together, they'll charge them for trafficking each other. "Anybody who's not a client who has something to do with sex workers can be called a trafficker," says McNeill. So, Thorn's narrow term for this ill-defined concept may be justified by a principled stance against exploitation, but it may also limit their ability to properly deal with situations that fall outside that framework.

Surveillance of what's out in the open

Other data sets led Thorn to sites Craigslist and Backpage, the latter of which circulates 45,000 escort listings every week.

Schmidt highlighted two ineffective approaches to dealing with sites known to traffic minors. One is through shutting down "adult service" pages, which some groups pressured Craigslist to do in 2010. This predictably resulted in ads simply migrating elsewhere. The second approach is through standard law enforcement practices, which are surprisingly outdated. Officers manually, tediously scroll and guess through pages of posts with potentially fake pictures and almost certainly fake ages.

Aware of the need to not violate the free speech of those advertising adult work, Thorn avoided the first approach and decided to improve the second.

They developed "an algorithm … that could help law enforcement surface the ads that were most likely to be children," said Schmidt. This algorithm can quickly scrape a webpage, analyze the semantics of wonky listings with "texting language" that younger people are more likely to use, "translate" obfuscated phone numbers, pick out key words advertising "youth," and put them all together in a spreadsheet. Tested for three months against previously confirmed posts of children, it was able to "identify patterns that wouldn't even be very obvious to the human eye," according to Schmidt. This helped them develop the algorithm to register an "underage likelihood score," and will continue build more "scores and indexes" as it scans more pages.

Again, there are limits. Thorn admits they don't have any tools to work on the surveillance-resistant Tor network. During a Q&A session after Stanford stopped recording the talk, Schmidt explained about the algorithm, "We don't want to say, 'We just caught you and here's how,'" because traffickers could then game the system.

Any development in police surveillance capabilities should warrant scrutiny, and it will be worth watching trafficking-related sting operations in cities like Anaheim, Houston, Minneapolis, Manhattan, Atlantic City, and Las Vegas, where law enforcement are already using Thorn's algorithm. But ideally this technology is neutral and will make the legitimate police work of protecting people more effective through more powerful analyses of publicly posted information.

Helping people help themselves

Perhaps Thorn's best approach is one that acknowledges how much ending the exploitation of minors relies on the individual, whether victims or perpetrators, to change his or her own behavior when given access to the right tools.

Although the National Human Trafficking Resource Center (NHTRC) has a hotline, Thorn found that "81 percent of [minor sex workers] never saw" the number. The same percentage of minor sex workers said they would have "wanted help to escape at some point during their trafficking situation." In addressing this information gap, they focused in on another statistic: 75 percent of those surveyed had a cellphone while being trafficked. To make the NHTRC more accessible and allow traffickees to get help on their own terms, Thorn developed a text shortcode, "BEFREE," which is easier to remember than a phone number and allows for less conspicuous communication than a call. Schmidt noted that her group is working on making information about shortcode more available through "online ads, text messages, [and] print ads in hotels"–the top three formats surveyed victims said they would have liked.

And Thorn isn't just trying to help victims.

"As opposed to arresting everybody who's ever downloaded a child pornography file," Schmidt said, she wants to deter people with targeted messages. Thorn sets up decoy files and sites that redirect to "people who are already out there looking for child porn to click and end up … on our landing pages" where they'll be located on a map (no more intrusive than geotargeted online ads) and given links to help resources without getting police involved. Over three years "we've gotten a million people to come to our landing pages so far… and we're seeing about… seven percent" of people clicking those links, Schmidt said, which "is a really high percentage in advertising terms."

Thorn can't be expected to solve all the problems of child sex trafficking. The organization is moving outreach in the right direction with its technological focus, but Thorn is not flawless. They're working on a broad problem that needs many tailored responses. Given the tech world's creative, competitive, and charitable culture, we may soon see even better ideas emerge on how to better protect the rights of these vulnerable members of society.  

NEXT: Sen. Chuck Schumer Blasts FDA's 'Ridiculous' New Proposal for Breweries

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. “Children and teens are trafficked for sex and exploited for pornography, and the web amplifies this trade.”

    Oh really? Prove it!

    I expect more from Reason than to promulgate evidence-free assertions from moral panic NGO’s and government agencies.

    1. Dozens, perhaps scores of police procedural television shows prove that it is a real problem.

      1. What is a real problem is the 96% of all child sexual exploitation cases are committed by people whom the child knows and trusts. 2001 “Commercial Sexual Exploitation of Children in the US, Canada and Mexico” page 92- 93 says that only 4% of all such exploitation is at the hands of strangers- not all of which is ‘sex trafficking.’ “The bulk of these assaults are perpetrated against children 12 years of age or younger and nearly all (84%) occur in the privacy of the child’s own home. Sadly, 96% of all child sexual assaults are perpetrated by persons known either to the child or the child’s family– 96% by acquaintances (e.g., neighbors, teachers, coaches, physicians), or by members of the child’s own family (e.g., fathers, step-fathers, uncles, older siblings). Contrary to widely held belief, only a small number of substantiated child sexual assault are committed by strangers (DoJ, 2000b:29)”

        1. Norma,

          I’ve heard of you, and greatly respect your work. Just wanted to mention that.


      2. What’s even worse is that a good part of the 96% of predators whom the child knows and trusts are law enforcement officers. And the children they tend to sexually exploit are far younger than the ‘children’ who are ‘sex trafficked’- where those ages tend to be around 16 and 17- whereas cops like very young children (they also like minors who are involved in the Explorer program- ages 10 to 17). And when COPS rape children, or download child porn, they tend to receive far less punishment than a non law enforcement agent pedophile. In 2013, Pike County Ohio Chief Deputy Clyde Franklin Sanders Jr. plead guilty to raping a THREE YEAR OLD GIRL- TWICE- and received probation. More of these horror stories can be found here:

        So I wouldn’t trust those police procedural television shows if I were you. Police departments receive huge grants from the Federal Government to find and ‘rescue’ those children.

      3. And when the FBI conducts it annual “Operation Cross Country” they have literally thousands of agents across the country looking for victims- and yet, they can barely find 2 or three victims in each city. In 2013, a total of 3,900 agents- federal, state and local- in 76 cities to rescue 105 ‘victims’ and arrest 150 ‘pimps.’ Either the pimps were time sharing those children or there are just not that many victims to be found. Each pimp would have the services of .7 children each- and it took 37.1 agents to rescue 1 child victim- and 1.4 children per city… this information comes from the FBI website- search for Operation Cross Country…
        http://www.policeprostitutiona…..l/Database Spreadsheets/2012/2012 pdfs/2011-13-Operation_Cross_Country_charts.pdf

      4. I assume this is sarcasm. If not, then I’m scared.

    2. I know- how is it that Reason can publish two great articles (Thaddeus Russell and Elizabeth Brown) which are well researched and then this one- which basically has all the same BS that the abolitionists spew out day after day! The number of ‘child sex trafficking’ victims is so small compared to the number of child sexual exploitation victims at the hands of cops, priests, preachers, teachers, coaches, parents, neighbors, etc. that it is amazing that so many people ignore those victims because there is no money involved in the transaction… perhaps it is because there is no funding available to crusade against the pillars of the community who are far more likely to be the predators than someone seeking to buy the services of children?

  2. “The National Center for Missing and Exploited Children received 500,000 potentially explicit images of minors online in 2004. By 2011 it surged to over 17 million.”

    First, how are these elements defined as “potentially explicit images of minors”?

    What’s with the words “potentially explicit,” anyway? Are they explicit – genuinely explicit – or not?

    Second, how old are these “images of minors”? Are we speaking of photographs and movies made in the 1950s and 1960s and 1970s, in which the “minors” pictured are now people in their fifties and sixties and even older? Or long since deceased?

    Moreover, how many of the counted millions of images are duplicates, logging identical copies hundreds and thousands of times over?

    It should be easy enough to specify “explicit” in the sexual sense, and to state without evasion how many of the images recently circulating had been made recently to figure in the exploitation of “minors” who are minors and whose best interests need to be served – urgently – by discovery and rescue.

    Are these thorny endeavors aimed at protecting the rights of victims, or punishing the thoughtcrimes of people who are doing nothing more exploitative or immoral than looking at photographs of train wrecks and battlefields?

    1. Yeah, that “potentially” is very misleading. Google, other search engines, and websites automatically report any flagged or reported image to the National Center for Missing and Exploited Children. This includes people who report images of girls who they think look younger than 18, as well as fictional images, not to mention things that turn out to be duplicates as mentioned.

      It’s also interesting how a change in law defined a moral panic literally overnight. The same act, legal a day before became a heinous crime the next day. When they set the drinking age to 21 in 1984, they also increased the age of consent from 16 to 18.

      1. Not to mention the images that we sex workers photoshop to make ourselves look younger- and also use photos of us when we WERE younger. If you saw my photos when I first starting working at age 32, you would have thought I was much younger, because I used photos of myself (for the madam I worked for) of when I WAS younger….

    2. Nudity, alone, is enough to qualify as “child porn” under current federal statutes (see Azov Films prosecutions, now unfolding). “Explicit” is ALWAYS alleged. And since it is illegal for anyone (other than law enforcement) to view the “child porn” in question, how would any of us know what the truth is?

      I am flabbergasted at the willingness of citizens, even so-called libertarians, to take allegations made by government agents at face value.

  3. Many of us sex workers are computer literate and know how to use photoshop to take years off our faces and bodies… perhaps many of those images were touched up to make the sex worker look much younger than she really is, because despite the millions of taxpayer dollars spent looking for child victims of sex trafficking, the number of arrests/rescues of minors for prostitution has steadily declined from 1981 to 2012 – in 1982, the police arrested 2,316 female minors.
    In 2012, the last year for stats from the FBI Bureau of Justice Statistics, there were 470 female minors arrested. The government tables which show the stats can be found on the FBI website, tables #40.

  4. This from the Washington Post: Lies, damned lies and sex work statistics…..tatistics/

Please to post comments

Comments are closed.