Can Anybody Find Me Somebody (Safe) To Love?
A procedural framework for screening convicted sex
offenders from online dating services
SUBMITTED IN PARTIAL FULLFILLMENT FOR THE DEGREE OF MASTER
OF SCIENCE
Miriam Soesan
11241047
M
ASTER
I
NFORMATION
S
TUDIES
H
UMAN-
C
ENTERED
M
ULTIMEDIA
F
ACULTY OF
S
CIENCE
U
NIVERSITY OF
A
MSTERDAM
July 20, 2018
1st Supervisor 2nd Supervisor
Dr. Frank Nack Prof. dr. Tom van Engers Informatics Institute Leibniz Centre for Law University of Amsterdam University of Amsterdam
Can Anybody Find Me Somebody (Safe) To Love?
A procedural framework for screening convicted sex offenders from online dating services
Miriam Soesan
University of Amsterdam Amsterdam, the Netherlands
ms.soesan@gmail.com
ABSTRACT
This paper describes a conceptual system that serves as a barrier for convicted sex offenders from using online dating services (ODSs). The foundational structure of the system lays in the connection and communication between the state, the ODS and the user. The system is designed to be implemented in the signing up stage of the user to the ODS. The system description and interaction design were focused around the notion of privacy and with respect to the Dutch law and the GDPR. By demonstrating the development of the proposed system, it becomes clear that a safer online user environment comes at the expense of anonymity and possibly at the expense of free online service.
KEYWORDS
Convicted sex offenders, registered sex offenders, online dating services, certificate of good behaviour (VOG).
1
INTRODUCTION
Online dating has become a popular tool for initiating romantic relationships. It is now one of the most frequently used services on the Internet, and one of its largest revenue generators. A piv-otal component of this phenomenon are mobile dating applications (apps), especially the ones with a geolocation function, enabling so-cial, romantic and sexual connections between proximate strangers [10, 20]. The multitude of platforms that facilitate the online con-nection and communication with strangers is predicted to continue to increase, with matchmaking apps reporting to have millions reg-istered users and billions of connections made1. In the Netherlands, it is estimated that more than 1.5 million adults use dating apps2.
Online services, and specifically mobile apps, provide new op-portunities for criminal activity, including sexual exploitation. The Internet is a largely unregulated environment in which it is possi-ble lure potential victims [3]. Routine activity theory posits that predatory criminal behavior requires motivated offenders, suitable targets, and lack of supervision [6]. Applying this theory to online predators, online services and the mobile app market cater both suitable targets and lack of supervision needed for motivated of-fenders. Dating apps can alter the opportunity structures within a user’s social ecologies, increasing their exposure to potential of-fenders [3, 5]. The inability to identify online predators creates a public threat [1].
Using ODSs to sexually offend is a global problem [1]. According to a 2016 report by the United Kingdom‘s National Crime Agency
1 https://www.datingsitesreviews.com/staticpages/index.php?page=Online-Dating-Industry-Facts-Statistics
2 https://www.nrc.nl/nieuws/2016/09/20/swipend-speuren-naar-ware-liefde-4374027-a1522306
(NCA), from 2009 to 2014 there has been a significant increase in reports to UK police forces regarding sexual assaults related to online dating [19]. Results of a recent study performed on students in Hong Kong have shown that users of dating apps were twice more likely to be sexually abused than non-users. While the study did not find indication of the apps to be the casualty leading to abuse, the authors found the association ’alarming’ [5].
Opening an online dating profile is usually an easy and quick process. Some ODSs, such as Tinder3and Bumble4offer users the option to link their account to their Facebook (FB) profile. Other ODSs, like Plenty of Fish (POF)5, Lexa6and Elite Dating7require
the user to fill in personal details manually, including profession, hobbies, relationship status, sexual orientation. By accepting the ODSs Privacy Policy and Terms of Use the user declares that he or she are aged 18 or older and do not have any conviction regarding sexual assaults.
In 2011, Match.com8announced they will start screening their
US members against the US public registry of sex offenders [13]. This was a result of a class-action lawsuit against the site, brought forward by a woman who had been raped by her date from the website. The offender had a previous record of sexual assault [13]. However, not all countries in which ODSs operate have public databases of convicted or registered sex offenders that the ODSs could use to screen their members. The research presented in this paper was conducted within the Dutch cultural and social environment. The Dutch government holds no public record of convicted sex offenders (CSO) and there is no registry of offenders. To date, research regarding the threat of CSOs, whether online or offline, has mainly focused on the threat they pose on children and youth [1]. Moreover, information systems (IS) literature has remained largely silent on this critical field of research.
This paper aims to fill some of the void in this field of knowl-edge by investigating: what are the requirements of a system that could enable screening convicted sex offenders from using ODSs in the Netherlands?
This paper explores a potential way to enhance the safety of ODS users by blocking ODSs from the use of CSOs, while coping with state laws, privacy concerns and possible social consequences. To answer the research question a procedural framework was de-veloped, while taking into consideration solutions to the stated problem found in other countries, the current safety measurements
3tinder.com 4https://bumble.com/ 5pof.com
6https://www.lexa.nl/ 7https://www.elitedating.nl/
8One of the most popular ODSs and now a member of Match Group, incorporating
ODSs provide to their users and relevant legislation and regulations in the Netherlands. The paper also addresses the moral and ethical drawbacks of the suggested procedure.
The rest of this paper is organized as follows: Section 2 provides a review of related work; Section 3 describes the conceptual frame-work and its components; Section 4 explains the interaction design, as relevant for each stakeholder involved in the process; The evalu-ation of the proposed framework is presented in section 5; Section 6 discusses different ethical aspects of the conceptual framework; Section 7 outlines the conclusions that were deducted from the research process; and last, section 8 depicts the limitations of the conducted research and suggestions for future research directions.
2
RELATED WORK
2.1
Sex offenders on the Internet
The ability to be anonymous on the Internet has been integral to the surfing experience for most Internet users. Although this anonymity has its benefits, especially from a privacy protection perspective, the ability to be unidentifiable on the Internet has cre-ated a new public threat: A virtual world in which sex offenders can hide and prey on their victims, and Internet users are unable to identify CSOs online [1]. Academic literature shows concern that the information disclosed on social networking sites is being used by sex offenders to identify potential victims [2, 14]. This concern is enhanced regarding online dating services, as they provide preda-tors an easier access to locate and connect with potential victims and seduce them into offline encounters [2, 5].
One of the main differences between online dating sites and other online social networks, is that once initial contact has been established, partners must determine whether to pursue other forms of communication outside the dating platform. Research has found that most users prefer to meet potential partners Face-to-Face (FtF) relatively short after the initial communication [22]. The different environment created online allows the dating process to speed up the relationship process. Trust is built up quickly, often with sexual messages and disclosures made before FtF meetings. This rapidly evolving connection leads users to perceive that they are already in a well-established relationship with someone they have just met online, and will not regard them as strangers, despite never having met them FtF [5, 19]. The increased sense of trust is evident through victims engaging in risky behaviors, such as agreeing to meet within a private residence [2, 5, 22].
Today, people can easily access the Internet via smartphones, and many dating apps now use the global positioning system (GPS). The GPS-based technology altered the way people search for poten-tial partners for a romantic or sexual relationship. It allows them to identify other users nearby, thus saving them time and effort and ultimately facilitating more dates. Moreover, people carry smart-phones with them at all times. thus, dating app users can easily access a pool of people anytime and anywhere. Another impact smartphones had on contemporary dating patterns is that people can download dating apps for free, while traditional dating websites often require subscription fees [5]. Alongside the benefits offered by the mobile dating apps, such as accelerate the accessibility a person has to others, the users of these apps increase their risk of meeting someone who is sexually coercive. In the past, it was
not easy for offenders to interact with potential victims who were not in their social network; nowadays, dating apps undoubtedly break this boundary, allowing people to connect easily with others, thus enabling online predators to interact with proximate potential victims [2, 5, 22].
There is a debate as to whether online offenders are a distinct group of sex offenders, or a typical sex offenders using a new tech-nology. Moreover, there is a debate regarding the exact definition of online offenders. It seems that online offenders are largely defined as those who commit all their crimes in the online sphere, without a FtF encounter with their victim(s). Offenders that use the internet to track potential victims in order to execute their plans offline, are referred to as Internet facilitating offenders [3].
2.2
Legislation concerning sex offenders
Up to date, fifteen different countries9, including USA, enforce sex
offenders’ registries, in an attempt to protect the public from the perceived threat posed by sexual offenders [26, 27]. Sex offender registration laws require convicted sex offenders to provide iden-tifying information to law enforcement agencies, which is then maintained in a centralized database. This database is used to both supervise known offenders, as well as to assist law enforcement in apprehending those who are suspected of committing new crimes [27]. The different attitude towards sex offenders’ legislation among countries lies on a broad spectrum. The USA lies on one end of the spectrum, with laws that are increasingly punitive and are driven by a community protection approach and public sentiment [21]. These include restrictions on where registered sex offenders (RSOs) may live, restrictions on employment options, and most recently prohibitions on accessing the Internet [17]. On the other end of the spectrum, Western Continental European countries (among which is the Netherlands) are the least punitive in their policies toward sex offenders. Out of the latter, France is the only country to hold a registry of sex offenders [21].
The USA’s Internet restrictions on RSOs requires members of that group to report their e-mail addresses and screen names to law enforcement officials (who sometimes include this information on sex offender registry listings), and forbid offenders from accessing social networking sites. To comply with these restrictions, RSOs are required to submit to periodic, unannounced examinations of their computer equipment and to install hardware or software systems on their computer to monitor their use. Furthermore, sex offenders must inform law enforcement authorities if they have access to the Internet at home or at their place of employment [17, 24].
The policies regarding RSOs Internet use allow the authorities in the USA to monitor the offenders’ activity online. The public sex offenders’ registry enables ODS to screen their members, as demonstrated by Match Group members (e.g. Tinder, POF). The public registry is also open for inquiries by private citizens, enabling them to verify someone’s past. In this case, RSOs activity on ODSs in the USA can be scrutinized by government authorities, the ODS and ODS members themselves.
As mentioned previously, the Netherlands currently holds no reg-istry of sex offenders. However, on January 1st, 2018, the Longterm
9Australia, Canada, Chile, France, India, Ireland, Japan, Kenya, Singapore, South Africa, South Korea, Taiwan, Trinidad and Tobago, UK and USA
Supervision Act10has become effective11. The purpose of this law
is that those under supervision of the state (tbs)12as well as sex and violent delinquents, will stay under supervision for as long as necessary, and will return into society under personal restrictions. In case of relapse behavior there is a possibility of immediate in-tervention. In this way it is possible, if necessary, to monitor this group for a lifelong term. At the conditional release (v.i.)13there is the common rule that the convict will refrain from illegal acts for at least one year. On top of this, special conditions can be imposed to work on the change of behavior or to limit the convict’s freedom. The law introduces a new measure for (former) tbs-convicts along with sex and severe violence delinquents. This measure enables to monitor an offender for as long as required. This independent measure is enforced by a judge’s decision, upon request of the pros-ecution. The conditions may include restrictions on the freedom of movement and/or influencing behavior. For example, a location or area ban, or prohibition of consuming alcohol or drugs. The law does not specifically state the attitude towards CSOs’ use of the Internet.
2.3
GDPR
The General Data Protection Regulation (GDPR) is the regulation in EU law on data protection and privacy for all individuals within the European Union (EU) and the European Economic Area (EEA) [25]. It also addresses the export of personal data outside the EU and EEA. The GDPR became enforceable on 25 May 2018, and influenced the design of the procedural framework presented in this paper.
The GDPR aims primarily to give control to citizens and residents over their personal data and to simplify the regulatory environment for international businesses by unifying the regulation within the EU.
The regulation contains provisions and requirements pertaining to the processing of personal identifiable information of individuals inside the EU, and applies to all enterprises, that are doing busi-ness with the EEA, regardless of their physical location. Busibusi-ness processes that handle personal data must be built with data pro-tection by design and by default, meaning that personal data must be stored using pseudonymisation or full anonymisation, and use the highest-possible privacy settings by default, so that the data is not available publicly without explicit consent of the subject, and cannot be used to identify a subject without additional information stored separately. No personal data may be processed unless it is done under a lawful basis specified by the regulation, or if the data controller or processor has received explicit, opt-in consent from the data subject. The data subject has the right to revoke this permission at any time.
A processor of personal data must clearly disclose any data collection, declare the lawful basis and purpose for data processing, how long data is being retained, and if it is being shared with any third-parties or outside of the EU. Data subjects have the right to request a portable copy of the data collected by a processor in a common format, and the right to have their data erased under
10Wet Langdurig Toezicht
11 https://www.dji.nl/nieuws/2017/wet-langdurig-toezicht-per-1-januari-2018-in-werking.aspx
12terbeschikkingstelling
13Voorwaardelijke invrijheidstelling
certain circumstances. Public authorities, and businesses whose core activities center around regular or systematic processing of personal data, are required to employ a data protection officer (DPO), who is responsible for managing compliance with the GDPR. Because the GDPR is a regulation, not a directive, it does not require national governments to pass any enabling legislation and is directly binding and applicable.
2.4
Current safety measurements on ODSs
On top of state legislation, the design of the system presented in this paper was influenced by the current manner ODSs function, their requirements of users and the safety measurements they practice to protect their users from unwanted experiences initiated by the use of the ODS platform. For this purpose, a total of five ODSs were investigated as part of this research. Three out of the ODSs investigated originate from the USA and operate on a global level-Bumble, POF and Tinder. The other two inspected platforms were established in the Netherlands, but operate also outside the Dutch borders- Lexa and Elite Dating.
All users who wish to open an ODS account are required to accept the service’s Privacy Policy and Terms of Use. Accepting these includes a declaration that the user is aged 18 years or older, and that he or she has no convictions for sexual assault offenses.
A common feature of all the ODSs investigated in this research, was that a user can open an account immediately, by following a few simple steps. There is no filtration of accounts at the first step. ODSs that are situated in the USA claim to run regular cross-checking of their users’ database against the public sex offender registry. In the Netherlands, there is no such registry so the ODSs rely on the user’s declaration and acceptance of the Privacy Policy (PP) and Terms of Use (ToU).
As mentioned in the introduction, some ODSs enable users to open an account by connecting the user’s FB profile. The infor-mation the ODS will receive from connecting the FB profile is thoroughly explained in the PP of the ODS. Among that informa-tion is the user’s friends, public posts, interests etc. Using FB as a mean to create a dating-profile, also enables the ODS algorithm to confirm (to some extent) the identity of the user. Profile certifica-tion is performed seamlessly, by granting the ODS provider access to applicant’s FB profile through FB API, which results in obtaining a ’certification badge’ [20]. Norcie et al. [20] found that online users believe that the amount of friends of a FB profile is a good indicator of the sincerity of the profile, and is a better measurement of authenticity than the date on which the profile was opened.
All ODSs that were examined in this research had issued safety guidelines as well as behavior guidelines (sometimes referred to as ’community guidelines’)14. These documents also include a call for users to report any unwanted behavior or violation of the safety measurements, as depicted in the guidelines, PP or ToU. More-over, all the ODSs included in the research promoted safety tips, regarding both online and off-line behavior of their users, as well as emergency telephone numbers.
14Bumble: bumble.com/en/guidelines Elite Dating: elitedating.nl/zekerheid Lexa: lexa.nl/pages/misc/charter POF:pof.com/safety.aspx
Tinder: gotinder.com/community-guidelines , gotinder.com/safety 3
A common feature in some of the ODSs researched is the double opt-in feature. This feature enables a conversation between two users only after both have ’approved’ of one another. Bumble has taken another safety measurement for the female users, on top of the double opt-in feature, only the female user can initiate the conversation. If both parties are female, then after both ’approved’ one another any of the two can start the conversation. But, if the other user does not respond within 24 hours, the ’match’ is canceled.
2.5
Deceptive Behavior On ODSs
Since the introduction of the Internet, public discourse of this venue for relationship formation included concerns about participants’ privacy and security. The perceived dangers of meeting strangers online included potential threats of identity theft [8, 11], sexual predators [1, 2, 14], cyber stalking [8, 11], and concerns about mis-representation [8, 10, 16]. Online mismis-representation can have real consequences for online dating participants, who are often search-ing for a long-term romantic partner. In other Computer-Mediated Communications (CMC) contexts such as chat rooms or online role-playing games, in which anonymity or pseudonymity are ac-ceptable or even expected. The anticipation of future FtF interaction inherent in online dating, provides motivation for reducing uncer-tainty in regards to the identity of potential relational partners through self-disclosure and other communication strategies [8, 10]. In online dating, deceptive behavior is common due to the ease in which personal profiles can be manipulated. ODS personal profiles are photograph and text-based; therefore, it is considerably easier for online daters to deceive others [10, 16, 20].
Erving Goffman [9] suggested that the packaging and editing of one’s self in the aim of making favorable impressions is an essential and ubiquitous component of social interaction. This process of self-presentation, however, is intrinsically intertwined with decep-tion, which is broadly defined as the intentional misrepresentation of information. Because individuals should be particularly con-scious of the impressions they create when constructing profiles to be scrutinized by potential mates, deception may be an impor-tant resource for creating an attractive self-presentation online [5]. However, when self-presentation lacks to present one’s dark past, as presented in Horcher’s [13] article it may impose a threat to the users of the dating website. One of the most difficult challenges in enhancing trust, security and privacy in most networked systems has always been, and remains, the human factor [20].
Walther’s [28] Hyperpersonal model is a useful framework for exploring how users take advantage of CMC deceptive potential, while considering the constraints of their relational goals. The model asserts that the distinct features of online communication offer users the opportunity to engage in selective selfpresentation -a more mindful -and str-ategic version of FtF self-present-ation. There are, however, social and technical aspects of CMC that may discour-age deception. For instance, recordability, or the ability to save and archive a profile, makes it difficult for users to erase evidence of their deception. Anticipation of future interaction, or users’ expec-tation of meeting other daters in person, where certain deceptions can be immediately spotted (e.g., height, weight, body type), is another deterrent against deception [10]. But, recordability and
future meetings may discourage those who wish to present a better version of themselves in hopes of gaining a romantic relationship. This may not be the case for users with malicious intentions.
2.6
Sex offenders’ social exclusion
With the aim to control crime and increase public safety in the USA, different monitoring and surveillance technologies have been utilized by correctional officials. Among these methods are track-ing and location systems, pharmacological interventions and risk assessment tools [17]. The most obvious form of technological inter-vention with regards to legislation has been the development of the ’magic bracelet’ and similar devices worn by paroled offenders. The bracelet consists of either radio or GPS-enabled tracking, allowing his or her movements to be tracked, either in real-time or as a set of aggregated data [29].
The cell phone, and later the smartphone, have played a role in the creation of a perceived need for localization systems that can be used by the general public [15, 17]. This has led to the development of sex offender’s alert software such as Sex Offender Tracker15, WeAlert16and Family Watchdog17. Such technologies may create
physical zones of inclusion and exclusion, typically enforcing the RSOs out.
Assessments of sex offender registration and notification policies are not widely prevalent, but to date such policies demonstrated only minimal effectiveness at reducing recidivism and victimiza-tions [27]. Residential restricvictimiza-tions have been shown to impose severe limitations on available housing options for RSOs [23, 24]. Employment restrictions in various jurisdictions include prohibi-tions on RSOs from working in schools, day care centers, and other (both paid and unpaid) positions where access to children is easy and common [23]. Each of these varieties of restrictions obviously presents potential limitations on social, employment, and other opportunities for RSOs and may have far-reaching individual and social consequences. Collateral consequences include relegating sex offenders to less desirable communities, relationship difficulties, heightened perceptions of vulnerability and stigmatization [17, 24]. In Tewksbury and Zgoba’s [24] research nearly all surveyed RSOs continued to use the Internet despite a state statute prohibiting such behavior. As reported by the RSOs who participated in this study, their primary online activities were focused on maintaining socially supportive contacts with family and friends, obtaining news, and seeking employment. These are pro-social activities that assist in the successful reintegration of offenders. However, by prohibiting such activities, there are likely to be weakened supports for CSOs and therefore fewer and weaker barriers to re-offending.
3
SYSTEM DESCRIPTION
This section provides a summary of the stakeholders, their role in the process and the data processing procedure of the theoretical framework developed. The framework is designed to operate by existing technologies and information systems. Therefore, some of the steps in the process should be familiar to the Dutch user, and are only mentioned by name.
15http://www.sexoffendertrackerapp.com/ 16https://wealert.io/
17https://www.familywatchdog.us/ 4
Figure 1: A step-by-step diagram of the suggested framework
3.1
Stakeholders
The framework developed during this research incorporates three stakeholders- the user, the ODS and the state, represented by the Dutch State Secretary for Justice and Security18. Each stake holder
has a different interaction with the system, a slightly different perspective and influences the flow of information in its turn.
18Ministerie van Justitie en Veiligheid
3.1.1 Justis. The Screening Authority of the Dutch State Secre-tary for Justice and Security, and the government body responsible of issuing one’s criminal record19. The Dutch version of a criminal
record is called ’Certificate of Good Behaviour’20, and is referred by its dutch acronym VOG. The VOG is a document that can be
19https://www.justis.nl/
20Verklaring Omtrent het Gedrag (VOG) 5
issued upon request of either a public or governmental entity or a private company, given the company is registered at the Dutch
Chamber of Commerce21.
3.1.2 The User.Initiates the conceptual screening process by applying to the ODS. The user must agree to the ToU and PP of the ODS for the process to start. If the user wishes to sign up with a made up user name, he or she will have to provide their true identity to the ODS on a separate step, before the ODS sends a request for a VOG to Justis.
3.1.3 ODS. Places the request for a VOG by the Justis. This step should happen automatically, after the user provides identifying details and agreed to the PP and ToU. The ODS is also responsible for keeping the privacy of users and deleting their data in case the account has not been activated within 90 days.
3.2
System requirements
Cavoukian [4] described seven principles of privacy by design. These principles aim to embed privacy and data protection through-out the entire life cycle of technologies, from the early design stage to their deployment, use and ultimate disposal. Some of the key principles relevant to this paper are the proactive and preventative approach; privacy by default of the system; and full functional-ity and securfunctional-ity. Due to the sensitive information handled in the process, these principals were taken into consideration in the con-ceptual framework presented in figure 1.
Another integral component of the process is trust, required by all parties involved. Murayama et al. [18] discussed how the Japan-ese concept of ‘anshin’- the emotional component of trust - can be considered when building technology systems. This enhances the choice to validate the user’s identity and lack of sexual criminal record, with the use of an already existing and familiar system. Only after completing this step successfully the user will be able to create an ODS account and use the service.
As described in section 2.4, some ODSs use a their users’ FB account as a safety measurement to validate their identity. This approach can have its drawbacks. For example, a user could set up a fake FB profile, link it to his or her ODS profile, and accredit it even though the information is false. Therefore, the system described here does not use FB as a mean to validate one’s identity.
3.3
System Overview
The process described hereunder is designed to act as a sort of barrier in the registration phase of ODSs.
The process is set into motion by the user, when he or she wants to sign up for an ODS. After either connecting a FB account or filling in details manually, the user must agree to the ODS’s PP and ToU. Once the user has agreed, the ODS would automatically send a request for a VOG on the user’s name. In case the user chooses a made up name to use on the profile, he or she would be requested to provide true identity details, in order to request the VOG. Either way, as part of the privacy policy and as mentioned in section 3.1.3, the ODS is responsible to keep the privacy of its users and those who request to open an account.
21Kamer van Koophandel
Once the signing up process has been initiated, the process must be completed within 90 days. This is because the request for a profile on the ODS is in fact a process that generates and stores personal data of the user. According to the GDPR, no organization is allowed to hold personal data for longer than necessary. Considering the different phases and possible setbacks of the system, 90 days seemed like a reasonable period of time. If the ODS does not receive a VOG within 90 days, the user data is deleted from the company’s database.
If foreigners would apply for an ODS account, the system will face difficulties assuring their clean record. In the Netherlands, immigrants are not requested to provide a criminal record from their country of origin, when applying for residence permit. The Immigration and Naturalisation Service (IND)22only requests a deceleration from the immigrant that he or she have not been convicted of any criminal activity. Only if later on this declaration turns out to be false, the residence permit is canceled. Therefore, if a recently immigrated user would apply for an ODS, it is most likely that he or she will receive a VOG and will be able to activate an account. This problem may be solved if either every immigrant to the Netherlands would provide a criminal record; or if ODSs will implement the procedural framework in a global level, so that once a user account is ’approved’ in one country, it can operate in any country.
4
INTERACTION DESIGN
This section describes in detail the interaction of the relevant stake holder at each stage of the process.
After the user agrees to the PP and ToU the application for a VOG is initiated seamlessly in the background, by the ODS. To apply for a digital VOG, the ODS must log in with eHerkenning23, a means by which companies and organizations can identify online. After the request for the VOG has been placed, the user must also be able to provide the following details:
• Registered as a resident in the Personal Record Database24 of the municipality where one lives.
• Has a citizen service number (BSN). • Has a user name and password for DigiD25. • Has an Internet connection and an e-mail address. • Has a payment account with a bank that is affiliated with
iDEAL26.
After the ODS prepares the application online, the user in ques-tion will receive an e-mail with a link and a request code, stating that there was a VOG asked for him or her. Via a link in this e-mail that user can log in to DigiD, where he or she completes the VOG application by confirming that his or her data has been correctly recorded (if that is the case). The rate of the electronic VOG appli-cation is €33.85 and must be paid via iDEAL before the processing of the application. The application will then be handled by Justis. Justis informs the person in writing about the processing of his or her application. 22https://ind.nl/ 23https://www.eherkenning.nl/ 24Basisregistratie Personen (BRP) 25https://www.digid.nl/ 26https://www.ideal.nl/ 6
After the request is placed, Justis will decide on behalf of the State Secretary of Security and Justice if the person will receive a VOG. If the query shows that the person has not committed any criminal offenses, Justis will take a decision within 4 weeks of receiving the application. If a VOG can be issued the ODS will receive it digitally directly from Justis.
If the investigation of Justis shows that the person has com-mitted criminal offenses, Justis will assess whether these offenses constitute an obstacle to the purpose for which the VOG was ap-plied for. For example, a taxi driver who applies for a VOG and has been convicted several times for driving under the influence of narcotics is unlikely to receive the statement. This also applies to an accountant who has been convicted of fraud. A bookkeeper who has been convicted of driving under the influence of narcotics has a different chance that he or she will receive the statement. In the case presented in this paper it is likely that on top of CSOs, those who have been convicted with domestic abuse, violence within a romantic relationship and other crimes of the sort, will also be refused a VOG for dating services. If the offender has committed criminal offenses, he or she will receive a decision within 8 weeks of receipt of the request.
If there is an intention to decline the request for VOG the in-teraction between the user and Justis is done in writing. In this case, the person in question receives a letter from Justis informing him or her about the intention to reject the application for VOG. The user can then respond by submitting a written appeal. Then Justis makes a final decision. The user can object to this decision by appealing to the administrative court. If the court rules in favor of the user then he or she is granted a VOG and will be able to open the account at the ODS.
Either way, utilizing this system would affect the interaction users had with ODSs so far, as signing up will no longer be free (even though the payment is not to the ODS) and the account will not be instantly activated.
After the user has an ODS profile, he or she will only have further interaction with the ODS. If the user wishes to open multiple online dating accounts, he or she can do so by applying to other ODSs. If the user already has a VOG for one ODS, the interaction with the system will change. The ODS will place a request by Justis for the user’s VOG, but since it is a VOG for the same purpose and it was already issued previously, the user should not pay the fee, or wait again for four weeks.
Another interaction that shifts after the user is granted a VOG, is the one between the ODS and Justis. If a user’s VOG status changes after opening an online dating profile, Justis should inform the relevant ODS(s) of this change in status. Currently, Justis only issues a VOG upon request. If a company or organization want an update on a person’s VOG status, they have to request it again through the same process.
5
EVALUATION
The theoretical framework developed in this research was evaluated by semi-structured interviews with experts. The research made use of convenience sampling, meaning the sample set has been generated by virtue of accessibility. The expert evaluators were contacted through personal social networks and through official
channels of the stakeholders concerned in the suggested system. In total, thirteen ODSs were contacted, three consumer protection organizations, five UX experts, three lawyer offices and the Dutch Ministry of Justis and Security.
The response rate was very low, and out of twenty five requests for evaluation, only three took place eventually. The first interview was conducted with a UX expert, and the other two were performed with two different ODSs representatives. The goal of the interviews was to gain an in-depth perspective of the stakeholders involved in the framework. For this purpose the participants were presented with the sequential diagram, and were walked through the steps the ODS applicant would have to take. Explanations regarding the choices made along the way were also provided to enhance the understanding of the final proposed process.
During the interviews the respondents had comments regarding the wordings chosen for the diagram. The wordings were changed later to make it more comprehensive.
The main concern of all three participants was that the suggested framework makes it harder for users to sign up for ODSs. Most ODSs users expect the service to be free of costs and to be able to use the service immediately upon registration. Having to pay for the service and wait to be approved before using the service will also make it harder for ODSs to run their business.
All three were focused on methods to reduce criminal activities through the ODS platform, rather than preventing the access of un-wanted users. For example, all three suggested to take an approach that would make it easier for users to report or review one another. In this way, users will be more intimidates to try and abuse the system. Two of the interviewees also suggested to add warning notifications if a user is ’matched’ or chatting with another user that has recently been reported for offensive behavior. Another suggestion was that ODSs will organize offline meet-ups for their users, to reduce anxiety and enhance trust towards the platform and its users.
The assessors were also concerned that the process to get a VOG should only be done in the background, and to involve the user as less as possible. One of them also suggested that instead of a digital VOG process, each user could go to the municipality office and request a VOG in hard copy. He or she could later send that VOG to any online or offline dating service, as they choose.
One interviewee also wondered if setting up such a system would actually stimulate users to commit sexual assaults. His claim was that this hurdle in the registration phase might bring the possibility of exploiting the dating platforms to users who otherwise would not have thought about it.
Some of the issues discussed during the evaluations were already implemented in the final version of the procedural framework. Other issues are discussed in the following section.
6
DISCUSSION
6.1
Anonymity
The Web as we know it today mostly facilitates free and anonymous usage. A system like this takes away both components in the aim of creating a safer environment for the general public. Users may prefer to be anonymous or use pseudonym to hide their identity
in ODS for different reasons. For example, a person who is embar-rassed of using such a service, or a person who uses the service to cheat on his or her spouse. The fact that the system obligates the users to confess their true identity might dismiss those who do not wish to reveal their identity, and not because of malicious intentions. On the other hand, this may also operate as a barrier for CSO who will not even try to sign up, as they know the process is bound to end with no ODS profile. Moreover, the payment issue may also act as a barrier for those who know that this is not a good investment for them.
6.2
Privacy
Opponents of online dating regulation may claim that ODS users should have the right to control what personal information they share, including felony incarcerations. The claim might be that it is not the state’s place to force users to disclose such private information. Additionally, the system that is presented in this paper might be viewed as one that is one step closer to a surveillance society, in which every action is investigated. This argument can be dismissed. Even though in the Netherlands criminal convictions are not public records, released CSO may still be defined as one that imposes danger to others, by a judge’s rule. If this is the case, then it is in public interest to bar CSO from platforms that might be used to find potential victims. Additionally, this kind of disclosure is no different than what companies and organizations already request on job applications.
Another privacy argument states that dating websites make the dating process safer than offline matchmaking platforms, such as newspaper singles advertisements and telephone-dating services, because online profiles supply users with more information than offline dating services. This enables users to better evaluate the person’s profile before talking on the phone or agreeing to meet.
But, online dating should be safer than meeting someone in a bar because bars merely provide a place for customers to eat and drink; meeting potential dates is only a byproduct of that service. ODSs’ primary service is to actively match their users. Moreover, if the argument is that ODSs are safer because they reveal more information about prospects, then these platforms websites should include meaningful information- more than income level, physical characteristics, and a photograph.
The GDPR does reinforce the privacy argument, as it limits the data that online platform companies can request from users, and the period of time that they can hold the data. Since criminal record is considered sensitive data it is important that ODSs will not hold that data themselves. Therefore the framework operates on a binary basis, in which the user either does or does not get the VOG. If the user has not completed the process within 90 days, the ODS should be obliged to erase the data of that user. Since there are different reasons for a user not to complete the process (e.g. lack of willingness to pay the fee to Justis, loss of interest etc.), there risk of ODSs creating their own CSO database would be dispelled.
6.3
False sense of security
Officially screening members on ODS may result in a greater safety risk as it might provide users with a false sense of security. For example, if the Justis database is not up-to-date and if there is
no continuous communication between the ODSs and Justis that informs the ODS in case a user has been convicted with a rele-vant assault. Moreover, users with other criminal background or with malicious intentions may still find their way on to dating platforms. For example, users with intentions, and possibly crimi-nal background, of fraud might use the platform to extort others. Additionally, the system does not secure users against first-time offenders. Screening out users according to their past criminal assaults means that people with no previous criminal background can still open a profile and use it to assault others. The system does not ’protect’ users from future assaults, but rather filters out users that have proven to pose a danger on others.
The anti-screening argument fails because it assumes that users do not currently hold a false sense of security and would only acquire one if the ODSs screens applicants. Is it worse to have a false sense of security or no security at all? Each year, Internet predators commit more than sixteen thousand abductions, one-hundred murders and thousands of rapes stemming from online services, including dating websites [7].
The false sense of security may even be worse when it comes to users who pay their ODS. Users naturally assume there is a differ-ence between paid and unpaid services, and the major differdiffer-ence should be security [13]. The assumption that if websites screen applicants, users would be more reckless in their dating activities is without evidence.
At the same time, users must always beware and conduct per-sonal investigations on his or her match. As mentioned before, the fact that a person has not committed a felony in the past does not indicate that person’s future intentions. Laws are in place to deter, catch, and convict criminals, yet crimes still occur. Therefore, users must always take responsibility for their safety.
6.4
Liability
This paper raises the question of the obligation both ODSs and the state have towards providing a safe environment for users. If they are indeed obliged who should determine what that obligation is and how it is fulfilled? Should each government look after its own citizens or should there be global guidelines, as online platforms operate (almost) without geographical boundaries? Too much regu-lation would repress the spirit of choice involving the online dating experience. Moreover, ODSs might avoid operating in countries with too much regulation, due to fear of liability stemming from user interaction.
Another argument opposing regulating ODS is the belief that such regulation would impose unnecessary burdens on the industry and would not improve user safety [13]. Moreover, there are privacy implications when a website requires users to disclose personal information such as past crimes. Therefore, some would argue that it is enough for websites to display safety tips and investigate suspicious activity and complaints.
6.5
ODSs as jurisdictions
Some of the aspects that were inspected were related to the com-munication between the ODS and its users. One aspect that was investigates involved enhancing the number of users reporting back to the ODS after an unsuccessful date. Almond et al. [2] stated that
one of the difficulties in reducing the activity of sex offenders in ODSs is that most users who were assaulted by their date whom they have met online, do not report back to the ODS where they got acquainted. If the offenders are not reported, the ODS has no way of banning or suspending their accounts. Therefore, the offender can keep using the ODS and assault other users. Two possible solutions where inspected with regards to this problem.
First, researching the current funnel through which users can interact and report back to the ODS and see if it can be improved in order to enhance users’ reports. This way, the users can directly report to the ODS if they have been attacked and the ODS can respond immediately, for example by temporary suspending the aggressor’s account until the matter is clarified. If it turns out the aggressor is guilty, the ODS can delete the account permanently and not let the user back to the platform. The problem with this approach is that it addresses the ODS as a separate jurisdiction from the state(s) in which it operates. If a person was attacked or assaulted by another person, he or she needs to report the incident to the police. The police can then investigate and if necessary take legal actions against the aggressor. Only after the police and the court of law have found the aggressor guilty, can the ODS decide if and how they want to punish the aggressor.
Another approach was suggesting a review system that would enable users on ODSs to rate and review users whom they have dated, and if necessary- warn other users about them. This approach can easily turn into a system that is used for personal revenge or simply distort reality according to users’ whims.
At one point in the research the system included a phase in which if the user got a letter of from Justis listing the reasons for not issuing a VOG, and the reasons had nothing to do with sexual assaults, the user could send that letter to the ODS who could then decide to allow the user to open an account. The problem with this stage was that it placed the ODS as a higher authority than Justis. Since the ODS is ultimately a service, it cannot overrule the decision of the state’s justice system. For this reason, the system presented in this paper does not include that phase.
6.6
Financial aspect
Opponents of regulating ODS may argue that this sort of legislation would unfairly target online businesses, forcing them to invest more financial resources than their offline counterparts. Most offline dating channels are unregulated and are not required to screen. This might be perceived as a financially discriminating process, forced on online platforms, but not offline ones. However, the process scheme suggested in this paper is derived from the process of issuing a VOG. The VOG can be requested and issued either digitally, or by hard copy at the municipality or directly at Justis27. Therefore, with a few minor adjustments, the framework proposed in this paper could work for both online and offline dating services. In the procedural framework suggested in this paper, not only ODSs will have to face excessive investment of time and money, but also their users. This might be perceived as an unwanted burden, and some might prefer to investigate their proposed matches on their own. This paper presents a framework that is based on the current process for issuing a VOG and the fact that currently the
27https://www.justis.nl/producten/vog/vog-aanvragen/index.aspx
Netherlands hold no public record of CSOs, or any other convicted criminals. It might be that this background screening process is too cost-ineffective to implement, but it is hard to determine that. Providing ODS access to state sex offender databases might be more time-efficient to the process and reduce the costs for the users.
Another issue regarding the cost of such a system is the extra manpower it will demand from the ODSs and Justis in implementing the process. Nevertheless, insufficient staff is not a persuasive reason to ignore basic safety owed to customers.
Criminal background screening could be added to already ex-isting screening procedures performed by ODSs. For instance, be-tween the years 2000 and 2010, the dating website eHarmony28has
rejected approximately one million people [13]. The reasons for rejecting applicants include applicant’s age is below 20, personal status such as applicants who are separated but still married, or applicants who have been married more than twice. Another filter for applicants, used by ODSs are personality surveys like the ones used by POF and Elite Dating. These questionnaires go beyond asking users to describe themselves and what they look for in a partner. The questions are aimed at reveling the user’s interests, hobbies, characteristics, habits and beliefs. Some ODSs allow users to assign different weights to different categories, according to what they hold important, and be matched accordingly. However, these questionnaires do not take into account criminal history.
6.7
Social Impact
Creating a barrier for ODSs may have social implications such as frustration due to exclusion, stigmatization, damage to self es-teem and loneliness, especially for those who will be screened out. Therefore, the screening should last in accordance with the judge’s rule, on individual cases. This way the system would operate as a supporting framework for protecting the public from those who are regarded as dangerous. Tewksbury and Zgoba [24] examined how RSOs experience, respond to, and attribute stress regarding sex offender registration and notification process and policies in the USA. Their findings showed significant levels of stress resulting from limiting or blocking Internet access and significant losses due to restrictions on Internet access. A 2016 survey showed that public attitudes towards sex offender rehabilitation in the Netherlands was more positive in comparison with seven other European countries participating in the survey [12]. The aim of the system is not to increase the stress levels of recovering CSO, but it might indeed cause that. Therefore, it is of utmost importance that the screening phase will work according to state law, and not according to ODSs requirements or perception of the threat the CSO may or may not pose on other ODS users.
7
CONCLUSIONS
Like many online activities, online dating can be a valuable re-source if used appropriately. While ODSs bring people together that might not otherwise meet, the outcome is not always positive. For these situations, user safety must remain a primary concern. The system proposed in this paper encourages both ODSs and legis-lators to support user safety by providing a shield from dangerous subscribers. However, this research and the system proposed in
28eharmony.com 9
it, promote questions regarding the current use of online services as free of financial costs for users, free of specific regulation per service sphere and the ability to use the service anonymously.
Furthermore, this research showed that the implementation of such a system is possible. All technology and laws reviewed and referred to in this paper already exist. The challenge would be to persuade both ODSs and state officials that it is within their interest to protect their users or citizens.
8
LIMITATIONS AND FUTURE WORK
The current study took a novel approach to solve a contemporary problem, while operating within the Dutch social, cultural and legal environment. The system presented is conceptually interesting, but is not without limitations. First, the system could not be imple-mented and tested in a real-world environment. The system did go through some iterations, that were a result of deeper understanding of the three stake holders, the way they operate, their needs and their limitations. Second, the response rate for evaluating the sys-tem was fairly low and lacked the presence of a law representative. Third, the literature on sex offenders operating through ODS is scarce. In addition, I could hardly find evidence of the problem in Dutch media, official statistical reports or academic literature.
An expansion of the present study would require collecting as-sessments from more ODSs representatives, criminal law represen-tatives and customer protection organizations. Further work could also be done in adjusting the present system to other practices, such as online financial services, in which the focus might be to screen users who have committed fraud or embezzling.
Finally, the present research constituted a cross-sectional anal-ysis of state law and online dating practices. Whether these two will cooperate in the sake of users’ safety is unclear; but it has been noted that ODSs have updated their PP in accordance with the GDPR, implying that they do have something to lose if they do not fall in line with regulations. Yet the attitude of ODSs towards local laws and regulations remains largely investigated. Another inter-esting avenue for future research is in the comparison of convicted sexual offenders’ usage of social platforms with first time offenders who use the same platforms.
REFERENCES
[1] Connie S. Albert and A. F. Salam. 2011. Protecting Children Online: Identifying Registered Sex Offenders Presence On The Internet And Consequent Online Social Behavior. Thirty Second International Conference on Information Systems, Shanghai(2011), 1–12.
[2] Louise Almond, Michelle Ann McManus, and Hannah Chatterton. 2017. In-ternet Facilitated Rape: A Multivariate Model of Offense Behavior. Journal of Interpersonal Violence(2017).
[3] Kelly M. Babchishin, R. Karl Hanson, and Chantal A. Hermann. 2011. The characteristics of online sex offenders: A meta-analysis. Sexual Abuse: Journal of Research and Treatment23, 1 (2011), 92–123.
[4] Ann Cavoukian. 2011. Privacy by design in law, policy and practice. A white paper for regulators, decision-makers and policy-makers(2011).
[5] E. P. H. Choi, J. Y. H. Wong, and D. Y. T. Fong. 2016. An Emerging Risk Factor of Sexual Abuse: The Use of Smartphone Dating Applications. Sexual Abuse: A Journal of Research and Treatment(2016).
[6] Lawrence E Cohen and Marcus Felson. 1979. Social change and crime rate trends: A routine activity approach. American sociological review (1979), 588–608. [7] Michael Fire, Roy Goldschmidt, and Yuval Elovici. 2014. Online social networks:
threats and solutions. IEEE Communications Surveys & Tutorials 16, 4 (2014), 2019–2036.
[8] Jennifer L. Gibbs, Nicole B. Ellison, and Chih Hui Lai. 2011. First comes love, then comes google: An investigation of uncertainty reduction strategies and self-disclosure in online dating. Communication Research 38, 1 (2011), 70–100.
[9] Erving Goffman. 1959. The Presentation of Self in. Butler, Bodies that Matter (1959).
[10] Jeffrey T. Hancock, Catalina Toma, and Nicole Ellison. 2007. The truth about lying in online dating profiles. Proceedings of the SIGCHI conference on Human factors in computing systems - CHI ’07(2007), 449.
[11] Nicola Henry and Anastasia Powell. 2016. Sexual Violence in the Digital Age. Social & Legal Studies25, 4 (2016), 397–418.
[12] Mechtild A. H¨oing, Romulus Petrina, Laurie Hare Duke, Birgit V¨ollm, and Bas Vogelvang. 2016. Community support for sex offender rehabilitation in Europe. European Journal of Criminology13, 4 (2016), 491–516.
[13] Maureen Horcher. 2011. World Wide Web of Love, Lies, and Legislation: Why Online Dating Websites Should Screen Members. J. Marshall J. Computer & Info. L.29 (2011), 251.
[14] Carolina A Klein. 2014. Digital and Divergent: Sexual Behaviors on the Internet. J Am Acad Psychiatry Law42, 4 (2014), 495–503.
[15] Jill S Levenson, Yolanda N Brannon, Timothy Fortney, and Juanita Baker. 2007. Public perceptions about sex offenders and community protection policies. Anal-yses of Social Issues and Public Policy7, 1 (2007), 137–161.
[16] Shao-Kang Lo, Ai-Yun Hsieh, and Yu-Ping Chiu. 2013. Contradictory deceptive behavior in online dating. Computers in Human Behavior 29, 4 (2013), 1755–1762. [17] Sharif Mowlabocus. 2016. fiY’all need to hide your kids, hide your wife’: Mobile applications, risk and sex offender databases. New Media and Society 18, 11 (2016), 2469–2484.
[18] Yuko Murayama, Natsuko Hikage, Yasuhiro Fujihara, and Carl Hauser. 2007. The structure of the sense of security, Anshin. In International Workshop on Critical Information Infrastructures Security. Springer, 83–93.
[19] National Crime Agency. 2016. Emerging new threat in online dating: Initial trends in internet dating-initiated serious sexual assaults. (2016). http://www. nationalcrimeagency.gov.uk
[20] Gregory Norcie, Emiliano De Cristofaro, and Victoria Bellotti. 2013. Bootstrap-ping trust in online dating: Social verification of online dating profiles. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelli-gence and Lecture Notes in Bioinformatics)7862 LNCS (2013), 149–163. [21] Michael Petrunik and Linda. Deutschmann. 2008. The exclusion-inclusion
spec-trum in state and community response to sex. International Journal of Offender Therapy and Comparative Criminology52, 5 (2008), 499–519.
[22] Artemio Ramirez, Erin M. (Bryant) Sumner, Christina Fleuriet, and Megan Cole. 2015. When online dating partners meet offline: The effect of modality switch-ing on relational communication between online daters. Journal of Computer-Mediated Communication20, 1 (2015), 99–114.
[23] T. ten Bensel and L. L. Sample. 2016. Social Inclusion Despite Exclusionary Sex Offense Laws: How Registered Citizens Cope With Loneliness. Criminal Justice Policy Review(2016).
[24] Richard Tewksbury and Kristen M. Zgoba. 2010. Perceptions and Coping With Punishment. International Journal of Offender Therapy and Comparative Crimi-nology54, 4 (2010), 537–551.
[25] The European Parliament And The Council of The European Union. 2016. REG-ULATION (EU) 2016/679. (2016). https://eur-lex.europa.eu/legal-content/EN/ TXT/PDF/?uri=CELEX:32016R0679
[26] U.S. Department of Justice. 2016. SMART SUMMARY 2016: Global Survey of Sex Offender Registration and Notification Systems. (2016). https://smart.gov/ pdfs/global-survey-2016-final.pdf
[27] James Vess, Andrew Day, Martine Powell, and Joe Graffam. 2014. International sex offender registration laws: research and evaluation issues based on a review of current scientific literature. Police Practice and Research 15, 4 (2014), 322–335. [28] Joseph B Walther. 1996. Computer-mediated communication: Impersonal,
inter-personal, and hyperpersonal interaction. Communication research 23, 1 (1996), 3–43.
[29] Dick Whitfield. 2001. The magic bracelet: technology and offender supervision. Winchester: Waterside Press.
