USENIX Security 2019论文集.pdf

USENIX

2019

需积分: 0 104 浏览量更新于2023-05-14 收藏 45.3MB PDF 举报

身份认证购VIP最低享 7 折!

30元优惠券

资源详情

资源推荐

Computer Security and Privacy in the Interactions Between Victim Service

Providers and Human Trafﬁcking Survivors

Christine Chen

Paul G. Allen School of Computer Science & Engineering

University of Washington

Nicola Dell

The Jacobs Institute

Cornell Tech

Franziska Roesner

Paul G. Allen School of Computer Science & Engineering

University of Washington

Abstract

A victim service provider, or VSP, is a crucial partner in a

human trafﬁcking survivor’s recovery. VSPs provide or con-

nect survivors to resources such as medical care, legal ser-

vices, employment opportunities, etc. In this work, we study

VSP-survivor interactions from a computer security and pri-

vacy perspective. Through 17 semi-structured interviews

with staff members at VSPs and survivors of trafﬁcking, we

surface the role technology plays in VSP-survivor interac-

tions as well as related computer security and privacy con-

cerns and mitigations. Our results highlight various tensions

that VSPs must balance, including building trust with their

clients (often by giving them as much autonomy as possible)

while attempting to guide their use of technology to mitigate

risks around revictimization. We conclude with concrete rec-

ommendations for computer security and privacy technolo-

gists who wish to partner with VSPs to support and empower

trafﬁcking survivors.

1 Introduction

Human trafﬁcking is a crime in which a perpetrator, or “traf-

ﬁcker”, preys on vulnerable individuals through atrocities

such as sexual exploitation, forced labor, or the removal of

organs [30]. As a conservative estimate, around 24.9 mil-

lion individuals worldwide are being exploited in this man-

ner [16]. Technology is playing an increasing role in this

ecosystem, from enabling trafﬁcking via online platforms

(e.g., [2, 18]) to aiding in the detection and halting of traf-

ﬁcking (e.g., [6, 25]).

In this work, we focus on a previously understudied role

that technology plays in the human trafﬁcking ecosystem:

technology in the interactions between trafﬁcking survivors

and organizations known as victim service providers, or

VSPs. VSPs exist to support their clients by providing re-

sources such as temporary shelter, help with employment

and legal issues, and mental health support. In this work,

we focus on VSPs providing resources to individuals who

are exiting or recovering from a trafﬁcking situation. These

resources are critical in protecting these individuals from for-

mer or future exploiters (“revictimization”).

Our research is driven by the following questions: How do

VSPs communicate and interact with their clients (trafﬁcking

survivors), and, particularly, what role does technology play

in that interaction? What are VSPs’ computer security con-

cerns and threat models, both for themselves and on behalf of

their clients? What technical (or non-technical) strategies do

they use to mitigate these concerns? And, ultimately, what

opportunities exist to better safeguard VSPs and their clients

from a computer security perspective?

To investigate these questions, we conducted a qualitative

interview study with 17 participants, including staff mem-

bers at VSPs and several trafﬁcking survivors. We analyzed

these interviews using thematic analysis common in qualita-

tive research. Our ﬁndings shed light on the general role of

technology in VSP-survivor interactions (Section 4.1), the

computer security concerns and threat models of VSPs and

their clients (Section 4.2), and the corresponding defenses,

where present (Section 4.3). We identify fundamental ten-

sions and challenges that must be taken into account by tech-

nologists who wish to improve VSP and client security and

privacy (Section 5).

At a high level, we ﬁnd that VSPs make technology-

related choices with the goals of protecting their clients from

revictimization and other harm. Speciﬁc instances of how

VSPs protect clients include helping clients lock down social

media accounts and enforcing shelter rules restricting photos

or social media posts (that may reveal the shelter’s location).

We also ﬁnd that, sometimes, the most effective means for

VSPs and their clients to interact are not the most conducive

to client safety. For example, despite the potential risk of

trafﬁcker-compromised accounts, some VSPs use Facebook

to communicate with clients because it provides a reliable

way to reach them even in the absence of cellular service.

More generally, we ﬁnd that our participants must balance

building client trust and maintaining contact with imposing

technology-related client safety rules.

From ﬁndings such as these, we distill concrete recom-

mendations for those in the computer security and privacy

community, and for technologists at large, wishing to help

support survivor-VSP relationships. For example, we pro-

vide guidelines on securing communications in situations

when the client’s device is compromised by an adversary

with physical access and raise awareness around the threat

posed to survivors by publicly available information (e.g.,

public records) online. In investigating the interactions be-

tween survivors of human trafﬁcking and VSPs from a com-

puter security and privacy perspective, our work contributes

to the larger push to leverage technology for good in the ﬁght

against human trafﬁcking.

2 Background and Related Work

There is a growing body of research examining the role of

technology in both facilitating and ﬁghting human trafﬁck-

ing (e.g., [2, 12, 17–19, 24]). In the computer science com-

munity in particular, prior work has developed technology

to aid investigators in examining online sex ads and online

forums for trafﬁcking activity [6, 25].

Focusing on the victim service provider ecosystem, there

has been research that explores the ways anti-trafﬁcking

organizations utilize technology to collaborate with each

other [29] as well as efforts within the VSP community

to leverage technology in providing help to trafﬁcking vic-

tims [20]. Work outside of the technical realm has examined

how survivors of trafﬁcking [4] or domestic violence [13]

experience and react to the assistance provided by VSPs.

Beyond human trafﬁcking, the computer security and pri-

vacy community has studied other speciﬁc (often at-risk)

populations, including journalists [22], refugees [28], and

undocumented immigrants [14]. Most relevant to our work

is research studying computer security and privacy for sur-

vivors of intimate partner violence [5, 10, 11, 21]. Where

relevant, we highlight similarities between our ﬁndings with

these prior studies on related populations.

In this work, we focus on an aspect of the human traf-

ﬁcking ecosystem that has not been rigorously studied from

an academic, technical perspective: the interactions between

VSPs and trafﬁcking survivors. We ask, from a computer

security and privacy perspective: how does technology en-

able or hinder these relationships, and how do VSPs and their

clients consider and mitigate the potential technology-related

risks that may undermine the survivor’s path to recovery?

3 Methodology

Between March and July 2018, we conducted 17 semi-

structured interviews with staff members at victim service

provider organizations and several trafﬁcking survivors.

Recruitment. Recruitment took place through several

primary methods: introductions facilitated by community

members and anti-trafﬁcking leaders, the authors’ personal

connections, and snowball sampling. Our recruitment adver-

tisements speciﬁed that we were looking for advocates who

work with labor trafﬁcking and/or sex trafﬁcking survivors

to speak about how they use technology in their work. We

also speciﬁed that participants would be compensated $30.

Participants. Table 1 provides an overview of the 17 study

participants. The 17 participants represented 11 different or-

ganizations; survivors P14 and P17 were not afﬁliated with

a speciﬁc organization at the time of the study. 16 partici-

pants were based in the U.S. and one participant was based

in a Southeast Asian country. Most participants were based

in urban areas.

As Table 1 shows, most participants currently focus on

serving survivors of sex trafﬁcking (though some partici-

pants may have previously helped labor trafﬁcking survivors

as well). To avoid confusion, we do our best throughout this

paper to call out results that are speciﬁc to interactions with

labor trafﬁcking survivors or sex trafﬁcking survivors. Fi-

nally, to be clear, note that some of the participants who fo-

cus on sex trafﬁcking survivors naturally also serve individ-

uals in the sex trade who may not technically fall within the

parameters of sex trafﬁcking (e.g. individuals who claim to

be in the sex trade voluntarily).

Study Protocol. The interviews ran between 60-90 min-

utes. We began with groundwork questions to understand the

participant’s role in supporting clients and general thoughts

on technology’s inﬂuence on the trafﬁcking ecosystem. We

then asked questions that would help surface how VSPs use

technology in their interactions with clients and what, if

any, concerns exist around this technology usage. We asked

about participants’ experiences with technology with regards

to ﬁrst contact with clients, client intake, organization and

client safety, and day-to-day interactions. To avoid priming

participants to overemphasize their computer security and

privacy concerns, most questions focused generally on tech-

nology in client-VSP interactions and related concerns but

did not mention computer security and privacy in particular.

Finally, we showed participants two prototypes for se-

cure communication (created by others): single-use URLs

(a URL that leads to sensitive content, which gets changed

to innocuous content when the same URL is accessed again)

and disappearing messages [1, 8, 9]. Our goal was to elicit

reactions and threat models using these concrete examples,

not to propose these particular technologies as perfect solu-

tions. To avoid participants giving inﬂated positive responses

towards the tools (participant response bias), we stated these

goals clearly for participants and also stated that we did not

make the tools. We asked questions like: When, if at all,

might you use this? How could it be helpful? How could

it introduce more risk? The full interview protocol can be

found in Appendix A.

ID Job Title Focus Client Nationality

Client Age

P1 Advocate, Survivor Leader Sex Trafﬁcking Domestic Adult

P2 Advocate Sex Trafﬁcking Domestic Adult

P3 Director Sex Trafﬁcking Domestic All

P4 Director Sex Trafﬁcking Domestic Youth, TAY

P5 Advocate Labor Trafﬁcking International Adult

P6 Director Sex Trafﬁcking Domestic All

P7 Advocate Sex Trafﬁcking Domestic Youth, TAY

P8 Advocate Labor and Sex Trafﬁcking Domestic, International All

P9 Advocate Sex Trafﬁcking Domestic Youth, TAY

P10 Advocate, Survivor Leader Sex Trafﬁcking Domestic Youth, TAY (to 25)

P11 Advocate Sex Trafﬁcking Domestic Youth, TAY (to 30)

P12 Advocate Labor and Sex Trafﬁcking International Adult

P13 Advocate Sex Trafﬁcking Domestic, International TAY

P14 Survivor Leader Sex Trafﬁcking N/A N/A

P15 Advocate Sex Trafﬁcking Domestic, International did not disclose

P16 Director Labor and Sex Trafﬁcking Domestic, International Adult

P17 Survivor Leader Sex Trafﬁcking N/A N/A

Table 1: Summary of Participants. Advocates support clients one-on-one, Directors oversee the VSP’s human trafﬁcking

services (managing advocates as well as interacting with clients), and Survivor Leaders are survivors of trafﬁcking (in this

case, sex trafﬁcking) who are raising awareness and leading trainings on the issue. Transition age youth (TAY) are individuals

between the ages of 16 and 24 [32]; where speciﬁed, participants also worked with clients slightly outside of this range.

Ethical Considerations. Our study was declared exempt by

the University of Washington human subjects review board

(IRB). We obtained informed consent from participants to

conduct and (optionally) to audio record the interview. As

the interviews could touch on highly sensitive topics (espe-

cially for survivors), we ensured that participants knew that

they could skip questions and request a break at any time.

We also emphasized that participants should provide only as

much detail in their answers as they felt comfortable with.

All electronic ﬁles were password protected, and physical

consent forms and notes were stored in a secure location.

Data Analysis. We continued conducting interviews until

no new themes emerged (saturation). We analyzed the data

thematically using a common methodology for qualitative

data [3]. We conducted multiple passes through the data

in which we iteratively identiﬁed and clustered themes, or

codes, present in the data. Two researchers independently

read through transcripts of several interviews, generated an

initial set of codes, met in person to develop an initial code-

book, and iteratively reﬁned this codebook by applying it to

additional interviews. Once the codebook was ﬁnalized, two

researchers divided up the remaining interviews and coded

them. We emphasize that the nature of our data is qualita-

tive, not quantitative, so we do not report on raw numbers of

participants who made certain statements in the results.

4 Results

We now turn to our results. After providing an overview

of the general practices our participants and their orga-

nizations use in interacting with trafﬁcking survivors, we

will present the security and privacy concerns and mitiga-

tion strategies — and tensions and challenges — that arise in

these interactions. We use the terms “survivor” and “client”

interchangeably, depending on the context and following the

norms described by our participants during the interviews.

At times, we also use the term “victim” and note that VSP

clients may not be fully removed from a trafﬁcking situation

when they are receiving services.

4.1 Client-VSP Interactions

This section provides background and context for the more

in-depth security and privacy discussions in later sections.

4.1.1 Role of VSPs

Though VSPs may help trafﬁcking victims escape their sit-

uations, their primary role is to help clients with the many

challenges they face on the path to stability, including look-

ing for employment, applying for housing, dealing with le-

gal matters, and coping with severe trauma. Importantly, as

we investigate in this paper, VSPs protect clients and train

clients to protect themselves from revictimization into a traf-

ﬁcking situation.

Some of our participants work at VSPs that provide shelter

for clients. These arrangements range from emergency shel-

ters (with very low barrier to entry — e.g., a client can stay

even if he or she is on drugs) to long-term homes (where the

client must be committed to actively working towards goals

and self-sustainability). As we discuss in Section 4.3, shelter

locations are sometimes conﬁdential to help protect clients.

As an overarching challenge in providing services to

clients, participants described the delicate balance they must

walk between building trust with their clients — so that they

can best advise and maintain contact with them — and doing

what they believe is best for the client. As clients have left (or

as they are in the process of deciding whether to leave) a sit-

uation where they have had little control over their lives, par-

ticipants often talked about how crucial it is to give clients as

much autonomy as possible. For example, P13 talked about

working with clients who want to ﬁnd a job. While she would

like her clients to go to school, she does not force her idea

of what would be best on the client. Throughout out results,

we will see this tension recur in the context of technology-

related guidelines and choices that VSPs are hesitant to push

on their clients.

4.1.2 First Contact

Clients typically make their ﬁrst contact with VSPs through

referrals — e.g., from law enforcement, schools, or other

VSPs — or via a phone hotline. Hotlines may receive emer-

gency calls, playing a similar role to 9-1-1 for clients and

trafﬁcking victims. For example, P7 described answering

hotline calls from individuals who are running for their lives

at the moment, and P8 talked about how they will dispatch a

Lyft or Uber to a caller who has just escaped.

Dispersal and discovery of hotline numbers happens in a

variety of ways. Beyond relying on word of mouth (a com-

mon method), participants talked about posters with the hot-

line number placed in public locations such as hospitals, train

stations, and rest stops. One organization has their hotline

number on a local Spanish TV channel. Another mode of

dispersal is through personal items (e.g., soap, essential oil,

hats, etc.) handed out to at-risk individuals (e.g. farm work-

ers) with the hotline number hidden discreetly on the object.

For individuals still in trafﬁcking situations, calling the

hotline can be dangerous (if the individual is constantly be-

ing monitored by their trafﬁcker) or even impossible (if the

individual does not have a device). In these situations, par-

ticipants described the ingenuity of their clients in ﬁnding

ways to access technology to get help. For example, one of

P16’s labor trafﬁcking clients saved up enough money from

tips to buy a burner phone from a gas station. While the

burner phone did not have the capability to connect to the

Internet, he had seen a hotline number earlier and committed

it to memory. As another example:

P8: I’ve had a few clients who, in escaping...

[were] able to get access to a hidden phone or

discretely (on an app that their trafﬁcker isn’t

aware...is a messaging app)...send messages to a

friend who helps them get help...

From advocates who work with sex trafﬁcking survivors,

we heard how clients will search the web for help:

P2: We’ve had a couple people. I’m like, “How did

you learn about us?” She goes, “I googled prosti-

tutes [city].”

At the same time, participants worried that lack of technical

expertise could make it challenging for clients to ﬁnd help

online. For example:

P1: I think what people have a hard time with is

search words. I think people don’t understand how

Google works, and how to search for things.

Mention of direct outreach by VSPs to potential clients

was rare, but one participant uses the phone numbers in on-

line sex ads to conduct text message campaigns to contact in-

dividuals who might want help leaving. Another participant,

P3, said that her organization reaches out to people who like

the organization’s Facebook page to see if they need help.

4.1.3 Continued Communication

Our participants typically communicate with clients via

phone calls, SMS, social media (e.g., Facebook), email, and

in person. Participants generally talked of using the com-

munication method that their clients feel most comfortable

with. P16 described how digital communication can help

put clients at ease.

P16: I ﬁnd that many of our clients are more com-

fortable engaging through technology because it’s

less raw. It’s a step removed in some ways...

Communication methods that work over WiFi were often

mentioned as important, as clients may not be able to afford

reliable cellular service or even a reliable device:

P2: A client right now has a phone. It’s not con-

nected to any service, but she can connect to WiFi,

so she and I can use Facebook Messenger instead

of texting. That’s true for a lot of our clients, be-

cause phones get turned off and on all the time,

numbers change all the time. I can still reach them

on Facebook, on Facebook Messenger. You can

As we will discuss further in Section 4.3, participants’ and

their clients’ threat models also inﬂuence their choice of

communication method.

4.2 Threat Models and Security Concerns

We now turn speciﬁcally to the threat models and computer

security related concerns voiced by our participants, both for

themselves as individuals and representatives of their organi-

zations, and on behalf of their clients. We found that many of

the security concerns or goals that our participants voiced ul-

timately revolved around preventing revictimization and pro-

tecting the physical safety of clients and VSPs. In this work,

we focus primarily on technology-related issues, but high-

light other concerns as well where necessary for context.

4.2.1 Trafﬁcker as Primary Adversary

The most common adversary for VSPs and clients were the

clients’ former trafﬁcker(s) or potential future trafﬁcker(s).

Compromising Online Accounts and Communications.

VSP clients’ communications may be compromised by traf-

ﬁckers, either digitally or via physical access. In many cases,

trafﬁckers have access to account credentials directly. P5,

who works with labor trafﬁcking survivors, described one

tactic trafﬁckers use to gain such access and alludes to the

way low digital literacy can harm international and/or labor

trafﬁcking victims:

P5: What if their trafﬁcker has access to their email

or helped them set up the email account. Just the

client never knew that and now I’m communicat-

ing with the client and [the trafﬁcker] is reading

our information?...I feel with our clients, they’re

just so vulnerable and a lot of them were brain-

washed...using a cell phone or using Facebook, a

lot of them, their trafﬁckers opened the account for

them and they think, “Oh he was just being helpful.

He wanted me to communicate with my family.”

Trafﬁckers may also compromise or intercept communica-

tions via physical access to clients’ devices. For example, in

the sex trafﬁcking ecosystem:

P7: I’ve had different guys that’ll pick up [my

clients’] phone and pretend to be them, go through

their messages.

Despite the risk that a trafﬁcker might physically see or

digitally intercept communication intended for a victim, P1

weighed such risks against the beneﬁts of reaching trafﬁck-

ing victims in her text outreach work. Note that the term

“pimp” is another way of referring to the trafﬁcker.

P1: And I don’t think it’s at the expense of the

victim, okay? I think, people ask this question be-

cause they’re like, “Well don’t you think that their

pimp is gonna beat them up because they got this

message?” Potentially. 100% yes...It’s either, I

get information out there that will potentially give

them an out, or they just don’t get anything.

Tracking Location. Another concern was trafﬁckers track-

ing down former victims after their escape. P16 has had

clients who found GPS trackers on their cars; P7 described

the use of tracking apps on phones:

P7: It’s usually...through [the victims’] device be-

cause most of the pimps get [the device], so they

have the family tracking, different apps and stuff

like that...One of my girls has shown me that they

can pull it up on their computer and you can see

where all of [the victims] are at one time.

Using Online Information to Track Down Survivors.

Even if a survivor’s devices or accounts are not directly com-

promised, participants worried about the use of online public

information to track down survivors. This fear is exacerbated

by the fact that the trafﬁcker often knows key information

(like birth date, social security number, etc.) that allows ac-

cess even to protected information.

For example, P14 is herself a survivor, and she general-

ized from her own experiences the ways trafﬁckers can uti-

lize public information to relocate survivors. Speciﬁcally,

she explained that trafﬁckers can ﬁnd where survivors have

moved by searching publicly available Department of Motor

Vehicles (DMV) records; they can use survivors’ addresses

and social security numbers to access and potentially lock

them out of their own bank accounts; and they can then track

survivors’ activities by observing the details of bank transac-

tions.

As another example, P17 described being found via med-

ical records and an old Facebook page she had thought was

gone. P16 talked about how shared rewards systems (like

grocery rewards cards) can reveal to a trafﬁcker where a sur-

vivor is shopping and what they are purchasing.

Undermining VSP Operations. Participants also discussed

the ways that trafﬁckers seek to undermine the efforts of

VSPs. P7 talked about trafﬁckers hanging out near VSPs

to recruit, and P16 talked about a trafﬁcker sending a victim

into a survivor program to recruit others directly. P2 men-

tioned that trafﬁckers have called her organization’s hotline

looking for survivors.

For shelters where the location is conﬁdential, participants

described various ways in which this conﬁdentiality could

be compromised. A common concern, for example, was that

people living in the shelter might accidentally reveal its loca-

tion (or the location of shelter guests) to trafﬁckers via pic-

tures or other posts on social media. P14 felt that location

conﬁdentiality was a challenging, if not impossible, goal:

P14: ...how conﬁdential really is any kind of build-

ing? I mean, you’re gonna see it on Google Maps

eventually. Whether or not you see it this year or

three years from now when they do their next pic-

ture, you’re gonna see it. So it’s not gonna be nec-

essarily conﬁdential for long.

剩余437页未读，继续阅读

阳阳爸

粉丝: 2
资源: 10

USENIX Security 2019论文集.pdf

USENIX2019.zip

USENIX Security2019.xlsx

四大网络安全顶会(NDSS S&P CCS USENIX Security)近三年（2016-2019）主会论文

2019 USENIX Security Symposium论文集：系统安全与物联网保障

"USENIX ATC 2019存储相关论文概述

Usenix Security Symposium

Usenix Security 2022

jabrefs 参考文献缩写

翻译这个网站下的内容https://www.usenix.org/system/files/sec23summer_254-nan-prepub.pdf

蓝牙配对漏洞 技术文档

2022年的安全安全相关学术讲座有哪些？

mapreduce相关文献

usenix伦理声明

usenix 2018

能列举一下USENIX ATC、HPCA、ASPLOS、OSDI、NSDI、EuroSys会议中有关操作系统的新算法吗，并详细讲解一下具体算法吗

能列举一下USENIX ATC、HPCA、ASPLOS、OSDI、NSDI、EuroSys会议中有关操作系统的新算法吗

什么是计算机三大顶会

给出分布式深度神经网络参考文献

tensorflow的外文文献

计算机毕设英文参考文献

最新资源

蓝牙配对漏洞技术文档