Assuming Identities Online: description, development and ethical implications.
Lead Research Organisation:
Aston University
Department Name: Sch of Languages and Social Sciences
Abstract
Preventive policing of serious crime sometimes involves deception and disguise. A case in point is the prevention of abuse arising from paedophile grooming and peer to peer networks where abuse images of children are discussed and exchanged. The preventive techniques by police investigators include assuming identities of existing community members, and of children, so that interventions and arrests can be made. Often, there are tight time constraints associated with this process - investigators have only a small window in which to learn and assume the identity in question before arousing suspicion in their target(s). The training that undercover online investigators currently receive, although broadly informed by linguistic theory, is in need of development. Furthermore, the time constraints mean that a semi-automated system to assist in identity assumption would represent a crucial contribution to the investigative toolkit.
Research taking a computational approach to the analysis of online communications has thus far focussed overwhelmingly on the structural elements of Computer Mediated Discourse (CMD), such as typography, orthography and other low level features, with little to no attention being paid to the socially situated discourses in which these features are embedded. The Centre for Forensic Linguistics (CFL) - a research centre within Aston University combining leading-edge research and investigative forensic practice - and Lexegesys - a consultancy and technology company specialising in developing and implementing data analysis solutions, recently collaborated on a project that was successful in automating the process of identification and extraction of low-level features for the purposes of attributing authorship of unknown texts within the context of Twitter. Yet CMD has widely been recognized to operate on a number of linguistic levels, such as those of meaning, of interaction, and of social practice. Outside of the computational linguistic field, the characteristic features of CMD are understood as resources that users draw on in the construction of identities in particular contexts, and CMD constitutes social practice in and of itself rather than simply being shaped by social variables.
Taking an inductive approach, which is to say that the phenomena of interest, rather than a specific theoretical paradigm, are primary, this research aims to bridge the gap between complex theories of the discursive construction of online identities on the one hand, and computational approaches to analysing online communications on the other. A small scale study CFL and Lexegesys are currently engaged in is addressing the challenges of automation at the pragmatic and interactional levels, working towards the semi-automated identification of phenomena such as indirect speech acts and topic management.
The work is extremely practical and is informed by real-world police investigations. A partner in the project, the West Midlands Police, Technical Intelligence Development Unit is crucially committed to providing data and operational insights. In addition to empirical applied linguistics, the project conducts proof-of-concept work for software that will assist in an ethical use of assumed identities in policing. Furthermore, it will involve an assessment of the ethical and policy implications for policing and security of complexity in online identity performance.
This proposal was previously submitted to the AHRC, and is resubmitted here on their advice.
Research taking a computational approach to the analysis of online communications has thus far focussed overwhelmingly on the structural elements of Computer Mediated Discourse (CMD), such as typography, orthography and other low level features, with little to no attention being paid to the socially situated discourses in which these features are embedded. The Centre for Forensic Linguistics (CFL) - a research centre within Aston University combining leading-edge research and investigative forensic practice - and Lexegesys - a consultancy and technology company specialising in developing and implementing data analysis solutions, recently collaborated on a project that was successful in automating the process of identification and extraction of low-level features for the purposes of attributing authorship of unknown texts within the context of Twitter. Yet CMD has widely been recognized to operate on a number of linguistic levels, such as those of meaning, of interaction, and of social practice. Outside of the computational linguistic field, the characteristic features of CMD are understood as resources that users draw on in the construction of identities in particular contexts, and CMD constitutes social practice in and of itself rather than simply being shaped by social variables.
Taking an inductive approach, which is to say that the phenomena of interest, rather than a specific theoretical paradigm, are primary, this research aims to bridge the gap between complex theories of the discursive construction of online identities on the one hand, and computational approaches to analysing online communications on the other. A small scale study CFL and Lexegesys are currently engaged in is addressing the challenges of automation at the pragmatic and interactional levels, working towards the semi-automated identification of phenomena such as indirect speech acts and topic management.
The work is extremely practical and is informed by real-world police investigations. A partner in the project, the West Midlands Police, Technical Intelligence Development Unit is crucially committed to providing data and operational insights. In addition to empirical applied linguistics, the project conducts proof-of-concept work for software that will assist in an ethical use of assumed identities in policing. Furthermore, it will involve an assessment of the ethical and policy implications for policing and security of complexity in online identity performance.
This proposal was previously submitted to the AHRC, and is resubmitted here on their advice.
Planned Impact
A major intended impact of the research is improved policing and a better protection of children online. The most significant beneficiaries of this research will be the children.
More direct beneficiaries include police officers engaged in online undercover work and specifically their work targeting paedophilic picture exchanges and online grooming activity. There are two separate aspects of impact within this task. The first is the development of policy concerning, and ethical conduct of investigations into, on-line paeodophile grooming, and the second is improved performance in the task, i.e. successful prosecutions of paedophile groomers. Different groups of police will benefit from the research project taken as a whole; locally, it will be the West Midlands Police (WMP) unit engaged in the project, but as this unit has representatives on the relevant committee of the Association of Chief Police Officers (ACPO) and holds the national training contract for online undercover operations, and as this training is increasingly available to international police, the impact will be rapidly disseminated nationally and potentially internationally.
In addition to this route for dissemination of impact the project outputs include the provision of individual training days and a training manual for the wider policing and security community and we would hope to attract to this training those with interests in investigation of high tech crimes, fraud, gangs and counter terrorism.
The research will also be of benefit to organisations such as the Child Exploitation and Online Protection Centre (CEOP) in its efforts to build up intelligence that informs their operational deployments and steers their educational programmes. Charitable organisations such as Barnardo's Cut Them Free campaign, which is working for strong and targeted policing to prevent sexual abuse and prosecute it when it occurs, will also benefit from the targeted preventative methods that will form the outcome of this research. The project also feeds into the work of the Virtual Global Taskforce, which seeks to build an effective, international partnership of law enforcement agencies, non government organisations and industry to help protect children from online abuse.
More direct beneficiaries include police officers engaged in online undercover work and specifically their work targeting paedophilic picture exchanges and online grooming activity. There are two separate aspects of impact within this task. The first is the development of policy concerning, and ethical conduct of investigations into, on-line paeodophile grooming, and the second is improved performance in the task, i.e. successful prosecutions of paedophile groomers. Different groups of police will benefit from the research project taken as a whole; locally, it will be the West Midlands Police (WMP) unit engaged in the project, but as this unit has representatives on the relevant committee of the Association of Chief Police Officers (ACPO) and holds the national training contract for online undercover operations, and as this training is increasingly available to international police, the impact will be rapidly disseminated nationally and potentially internationally.
In addition to this route for dissemination of impact the project outputs include the provision of individual training days and a training manual for the wider policing and security community and we would hope to attract to this training those with interests in investigation of high tech crimes, fraud, gangs and counter terrorism.
The research will also be of benefit to organisations such as the Child Exploitation and Online Protection Centre (CEOP) in its efforts to build up intelligence that informs their operational deployments and steers their educational programmes. Charitable organisations such as Barnardo's Cut Them Free campaign, which is working for strong and targeted policing to prevent sexual abuse and prosecute it when it occurs, will also benefit from the targeted preventative methods that will form the outcome of this research. The project also feeds into the work of the Virtual Global Taskforce, which seeks to build an effective, international partnership of law enforcement agencies, non government organisations and industry to help protect children from online abuse.
Publications
Grant T
(2016)
Assuming Identities Online: Experimental Linguistics Applied to the Policing of Online Paedophile Activity
in Applied Linguistics
MacLeod N
(2017)
"go on cam but dnt be dirty": linguistic levels of identity assumption in undercover online operations against child sex abusers.
in Language and Law / Linguagem e Direito
MacLeod N
(2016)
"You have ruined this entire experimentshall we stop talking now?" Orientations to the experimental setting as an interactional resource
in Discourse, Context & Media
Nathan C
(2016)
Liability to Deception and Manipulation: The Ethics of Undercover Policing
in Journal of Applied Philosophy
Sorell T
(2016)
Online Grooming and Preventive Justice
in Criminal Law and Philosophy
Sorell, T E
Violation of Privacy and Law: the Case of Stalking
in Law, Ethics and Philosophy
| Description | This project has generated a substantial amount of new knowledge in the area of linguistic identity disguise, as well as developing new research methods, identifying new research resources and opening up new research questions. These four achievements will be commented on briefly here. In a move rarely seen within the field of discourse analysis, where there has long existed a preference for 'naturally occurring' language data, a series of experiments were designed within this project. The objective of these experiments was to elicit Instant Messaging (IM) data which was a) content-neutral and b) easily controlled. The experimental data was considered a necessary bolster to the existing datasets that we had been provided through out connections with West Midlands Police (WMP), which comprised the following: 'genuine' IM data from conversations between offenders and victims in resolved cases of online grooming; 'operational' IM data from conversations between offenders and Undercover Officers (UCOs) in online operations against paedophilic file sharers; and 'training' IM data from conversations between trainee UCOs (as the victim) and their trainers (as the offender) in a roleplaying scenario concerning online grooming of a child. Though these datasets had the potential to be exploited for addressing some very important linguistic questions, the sensitive content tends to be unsuitable for publication, and they throw up numerous other complication (for an overview of the limitations of the datasets, see Grant and MacLeod 2016). The collection of experimental data sought to address these, and succeeded. Results from analysing the training data suggest a positive effect of input from forensic linguists on trainee UCO's performance in online identity assumption (see MacLeod & Grant, in prep.). While they continued to make some assumptions about their teenage target persona's language use based on her age and gender, linguistic input appeared to alert them to the dangers of stereotyping, and they demonstrated a greater sensitivity to her genuine evidenced linguistic choices after training. Analysis of the experimental data demonstrated that while individuals are adept at identifying, and to some degree emulating, certain structural features in a target's style that differ markedly from their own, they often neglect to sufficiently obscure marked choices in their own style. Thus, a participant's impersonation of another individual will represent a hybrid of the two individuals' usual styles. An analysis of comments made by individuals who were tasked with judging when their IM interlocutor had been substituted with another individual reveal that features at the structural level - spelling, punctuation, capitalisation and so on - are the most important in ascertaining the possibility of a switch. This is followed by features at the level of meaning, namely the proportional use of particular speech acts, and finally the interactional level, namely topic management. We continue to look into ways in which the fourth level of analysis identified by Herring (2004: 321), that of social behaviour, can easily be drawn on in real time operational contexts. With regard to the ethics and policy portion of the project we conclude in general that the construction of online personas, though deceptive, is overwhelmingly justified by the harm it prevents. Several underrecognised ethical issues come to light as a result of the project. First, there is a moral difference between taking over the identity of a real child who has been groomed, and inventing a child persona to attract grooming. The former is morally better than the latter, although both may contribute to prosecutions against dangerous people. The difference is that in one case but not the other the grooming would not have taken place without the creation of the persona. Second, where there is an overwhelming weight of beneficial consequences, it may be permissible for UCOs to act in ways that verge on entrapment for intelligence or preventive purposes, even where this is not aimed at a prosecution or where it is known that such activity would cause a prosecution to fail. Third, police acting undercover online may similarly omit to prevent harm where prevention would entail 'showing out' and thereby diminishing a unit's broader ability to prevent or apprehend a criminal group's activity. Regarding the legislative framework, we find, first, that the approach to grooming in section 15 of the Sexual Offences Act (2003) in particular is justifiable. It effectively distinguishes between "contact" groomers and "fantasy" groomers, removing a basis for doubt about grooming offences on the possible remoteness of grooming behaviour from actual sexual contact. Furthermore, it distinguishes between the sexualization of online behaviour between children and the informal profiling of likely targets by forensically aware adult paedophiles, who sometimes pose as sexualized children. Second, we find that there is a disconnect between the laws and practice surrounding intrusions by undercover police upon those who are relevant to an investigation but are not believed to be involved in a group's wrongdoing: although the Regulation of Investigatory Powers Act (2000) does not provide sufficient direct recognition of the higher ethical bar needed for intrusion on those who are innocent or only at the fringes of an organised criminal group, actual police practice surrounding covert strategies seems to express sensitivity to this issue. This is of especial significance in the online context, given the lower cost of making deployments. New research resources and research questions that we have identified for potential future work are discussed in the next section. |
| Exploitation Route | We hope to shift the focus of the analysis in future work to the language patterns associated with particular online criminal groups, with a view to assisting undercover operatives with successful infiltration of these groups. The concept of 'community of practice' (CoPs) (Lave and Wenger, 1991) is particularly useful in this context, and will represent a theoretically rigorous lens through which to view the groups under study. Membership of CoPs requires familiarity with group norms, including those surrounding communication (Eckert, 2006). This shift in focus will also necessitate a movement to different criminal contexts, since online grooming tends to be a relatively solitary practice. Other crime types, such as the sharing of indecent images, terrorism, and gang activity, tend to be more collaborative. It is hoped the impact of our work will then extend beyond the police service, potentially to the security services. Our research questions moving forward are: 1. What levels of linguistic analysis are necessary and sufficient to describe the linguistic character of an online community of practice in relation to the individuals in that community? 2. How much do criminal communities of practice vary in terms of linguistic coherence/ diversity and what best explains patterns of coherence and diversity across different communities? 3. How can a better understanding of the linguistic character of an online community of practice assist in the policing and infiltration of online criminal activity? |
| Sectors | Digital/Communication/Information Technologies (including Software),Government, Democracy and Justice,Security and Diplomacy |
| Description | The linguistic analysis of online personae has fed directly into the development of the IDentik software tool for assisting undercover online operatives with impersonation tasks. The first cohort of operatives to use the tool were trained in its use in March 2016. Since then a number of policing and security cohorts have been trained. As a result of the success of the software we are also setting up a spinoff company to provide a platform to supply the software into the police. Aside from the software tool, the research conducted within this project has also fed in to the linguistics training that the Centre for Forensic Linguistics (CFL) provides to the Pilgrim course. Set out in HMIC 2014, Pilgrim is a combined residential and distance learning course provided by West Midlands Police (WMP) to undercover officers seeking to move into online policing work. The content delivered by CFL covers the levels of linguistic analysis that can be useful in describing an online persona to the extent that an operative can assume that persona, including the levels of structure, meaning and interaction. The current project has strengthened CFL's ties with WMP and fostered fruitful relationships with the wider England and Wales policing movement. Reference HMIC (2014) An inspection of undercover policing in England and Wales. Available: https://www.justiceinspectorates.gov.uk/hmic/wp-content/uploads/an-inspection-of-undercover-policing-in-england-and-wales.pdf |
| First Year Of Impact | 2016 |
| Sector | Government, Democracy and Justice |
| Impact Types | Economic |
| Description | Ethics input for West Midlands Police training (ongoing) |
| Geographic Reach | National |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Description | Pilgrim training for online undercover police officers |
| Geographic Reach | National |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Impact | Delivery of linguistic training to online undercover police. Reference letters show change in policy and practice. |
| Title | UG and PG experimental chatlogs |
| Description | This data collection consists of transcripts of Instant Messaging conversations between a 'Judge' and an 'Interlocutor', the latter being replaced at some point by an 'Impersonator'. 3 x 15 minute chats per file, representing 3 conditions of preparation for the Impersonator in each case - No Preparation, Over the Shoulder preparation, and Homework preparation. The transcripts correspond to postgraduate students (files 1-12) and undergraduate students (files 13-30). Judges were asked to record when they thought a switch had taken place, what linguistic criteria led them to think this, and how confident they were in their decision. Information on when switches actually occurred was also collected, and cross referenced with these judgements. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2016 |
| Provided To Others? | Yes |
| Impact | In spite of the preference within a range of discourse analytical paradigms for 'naturally occurring' data, we argue that not only does the term prove conceptually problematic, but in certain contexts, and particularly in the applied forensic context, a rejection of experimentally elicited data would limit the possible types and extent of analyses. Thus, it would restrict the contribution that academic linguistics can make in addressing a serious social problem. There are multiple justifications for the use of experimental methods. Our 'situated' approach to identity might be seen to preclude the use of elicited or 'contrived' data in tackling our research questions, but the use of such data will allow for analyses to assist with the complex challenges thrown up by the project. Our research design demonstrates that superficial incongruity between one particular theoretical approach and the experimental method need not present an obstacle to conducting meaningful research. |
| URL | http://reshare.ukdataservice.ac.uk/852099/ |
| Description | 2LXG |
| Organisation | 2LXG |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | The Centre for Forensic Linguistics were responsible for carrying out the linguistic analysis underpinning this product. CFL met regularly with 2LXG throughout the project in order to feed back on the software as it was under development, set up trials of early versions, and offer input into the design and usability of the tool. |
| Collaborator Contribution | 2LXG, a local limited company specialising in engineering research and development, led work package 3 which was to produce a semi-automated system for (i) the analysis of online identities and (ii) to inform training in the area of undercover online investigations. The product of this work package is IDentik, a software tool which models a linguistic identity based on historical Instant Messaging (IM) data loaded by the user, and then provides real-time assistance as the user engages in an impersonation task over IM. |
| Impact | IDentik software tool. The tool will be offered at a small cost to online undercover officers nationwide. This is a multi-disciplinary collaboration between the Centre for Forensic Linguistics (CFL) and 2LXG, a company specialising in software research,. design and engineering. |
| Start Year | 2014 |
| Description | Assessor panel, CREST |
| Organisation | Centre for Research and Evidence on Security Threats |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Assessor of applications and member of commissioning panel. |
| Collaborator Contribution | The Centre for Research and Evidence on Security Threats (CREST) is a national hub for understanding, countering and mitigating security threats. CREST brings together the UK's foremost expertise in understanding the psychological and social drivers of the threat, the skills and technologies that enable its effective investigation, and the protective security measures that help counter the threat in the first place. It does so within a context of significant stakeholder and international researcher engagement, and with a clear plan for sustained and long-term growth. |
| Impact | Assessment of applications and drafting of calls |
| Start Year | 2017 |
| Description | West Midlands Police (WMP) |
| Organisation | West Midlands Police |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | CFL have been involved in training online undercover officers (UCOs) with the Technical Intelligence Development Unit at WMP for around five years, contributing input in the areas of linguistic modelling of online personae and the linguistic aspects of online impersonation. This has continued throughout the current project. The software tool created within the current project will be made available to UCOs nationwide through WMP. |
| Collaborator Contribution | WMP allowed CFL access to trainee online UCOs to use as research participants in the experimental tasks. WMP have also allowed CFL and 2LXG to observe trainees preparing for and participating in the impersonation task, which has allowed for the software to be designed with maximal consideration of the operational context. Lastly, WMP are facilitating the first IDentik trialling and training session in March 2016. This session will include the collection of feedback about the tool, with a view to revisiting the participants in three months to gauge how the tool is being put to use. |
| Impact | Pilgrim training course IDentik training course |
| Start Year | 2010 |
| Title | IDentik software tool set |
| Description | Software tool set to assist police in online undercover operations |
| IP Reference | |
| Protection | Copyrighted (e.g. software) |
| Year Protection Granted | 2016 |
| Licensed | Yes |
| Impact | IP licensed to development company for joint distribution and training. |
| Title | IDentik |
| Description | IDentik is a software tool designed to assist undercover operatives with impersonation tasks in the context of Instant Messaging (IM). Once loaded with a historical linguistic persona, the tool generates a model of this persona, and provides the operative with a wide range of information about it. These include an overview of variant and non-variant spellings, a list of spelling abbreviation rules, including frequency rates, and a 'persona' window which allows for the direct 'translation' of the operative's language into that of the target persona. |
| Type Of Technology | Software |
| Year Produced | 2016 |
| Impact | The IDentik tool is currently being trialled with undercover officers, from whom we intend to collect feedback after they have had access to it for three months. This should give us a clearer idea of the concrete impact of the tool on their professional practice. It is hoped, however, that the tool will relieve a substantial portion of the cognitive load associated with successfully performing impersonation, allowing for a greater amount of preparation time to be spent on other operational considerations. It is also intended to alleviate the problems associated with multiple operatives being required to impersonate the same individual, and equally with a single operative being required to impersonate several personae within a fairly limited stretch of time. |
| Description | Assuming identities online: software training workshop for online undercover investigators |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | Taking place in March 2016, this half day workshop took a subset of attendees at the National Online Undercover Working Group's annual CPD event (see separate entry), and provided them with initial training in linguistic analysis and impersonation using the IDentik software tool, and with a copy of the Beta version of the software itself. The tool was provided on an encrypted data key and assigned to each participant using their unique national ID number. Feedback was collected from participants via a focus group and questionnaire at the end of the session. We will replicate these feedback questionnaires with the same participants in 2-3 months time, when they have had the opportunity to put it to work in operational contexts. Both sets of feedback will be fed back to the developers of the tool and considered in future modifications of the tool in an ongoing process of development. |
| Year(s) Of Engagement Activity | 2016 |
| Description | BAAL conference - panel discussion |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | The role of descriptive linguistics in the policing of online paedophile activity. This paper is concerned with setting out our theoretical approach to the relationship between language and identity, with a strong focus on the practical application of this theory in the context of online identity assumption by undercover investigators in grooming cases. We describe a novel project within which we address the question of what linguistic analysis is necessary and sufficient to describe an online persona to the extent it could be successfully assumed by another individual. A central practical concern of the project is how we can best assist trainee investigators in moving away from a reliance on ideological assumptions about how particular social groups use language, towards a more theoretically informed approach that takes into account the complexities of situated identity performance. This is seemingly at odds with the approach to identity dominant within the field of author profiling, where the rather simplistic focus is on correlating particular linguistic categories with particular social categories. Through police partners on the project we have access to naturally occurring data in the form of real child abuse conversations and also data which arises out of the training of officers who investigate such offences. However, working with such data is difficult (in many different ways) and it is not wholly necessary to work with it to answer our more general research questions. We therefore argue that despite the privileged status awarded by sociolinguists and discourse analysts to 'authentic' or 'naturally-occurring' language data, experimental methods can in fact be invaluable for addressing theoretical questions of language variation and identity in this applied context. By shaking off methodological and theoretical constraints associated with our field of study, we are able to further enhance the contribution that linguistics can make in addressing a serious social problem. |
| Year(s) Of Engagement Activity | 2015 |
| URL | http://www.aston.ac.uk/lss/news/events/baal-annual-meeting/ |
| Description | BAAL/ Rouledge Research Development Workshop, Nottingham Trent University |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Around 30 Postgraduates and ECRs attended this event, in which we presented 'Assuming identities online: Applying sociolinguistics to undercover online policing'. |
| Year(s) Of Engagement Activity | 2016 |
| URL | http://www.baal.org.uk/workshops_2016_nottingham_programme.pdf |
| Description | CEPOL conference - Budapest |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Assuming Identities Online: Linguistic Contributions to Training Undercover Investigators. The issue of identity and influence within transnational online communities has become a significant policing and security concern. Online communications pose their own unique set of challenges to investigators of paedophilia, terrorism, organised crime and other threats. The anonymity offered by the internet has the potential to hamper policing, inevitably putting greater emphasis on forensic linguistic analysis for the tackling of online crime (Hughes et al., 2008). One tactic to address these issues can be to deploy undercover officers to take over the online accounts of either victims of crime or where possible of arrested offenders. With reference to the domains of Computer Mediated Discourse (CMD) set out by Herring (2004; 2014), we use analyses of online interactions to develop a linguistic model that can be used to improve performance in identity disguise and we demonstrate the domains of language that contribute to assessments of online personae as deceptive. We report on an evaluation of linguistic training delivered to UK online undercover officers, in the context of anti-paedophile operations, and demonstrate how such training assists in account takeovers, by improving officers' capabilities and reducing the chance of the impersonation being detected. Further to discussion of the linguistic training, we introduce the IDentik software tool which is being developed to assist officers in this task, and briefly show how it can be used to achieve analyse an online account to enable its successful takeover and thus support the officers in textual interactions in the course of an operation. |
| Year(s) Of Engagement Activity | 2016 |
| URL | https://www.cepol.europa.eu/ |
| Description | Consultations on Communications Data Bill - Home Office |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Attendance at Home Office consulting on the draft Investigatory Powers Bill |
| Year(s) Of Engagement Activity | 2015 |
| Description | Cybersecurity meetings - Washington |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Roundtable of thought leaders at the Washington campus of Johns Hopkins, plus large meeting at british embassy in February. Beginning of long term collaboration with Georgetown and Washington |
| Year(s) Of Engagement Activity | 2015 |
| Description | Ethical Issues in Undercover Policing (College of Policing commissioned special consultation) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | Special meeting funded by the College of Policing, organised by IERG at Warwick University. This meeting aims to enhance police understanding of the views of people with experience and relevant expertise on how undercover policing should be delivered and managed. It brings together a series of non-police experts in the field. A report of the event is to be submitted to the National Undercover Scrutiny Panel, an independent body established by the College of Policing to challenge police decision-making. The Scrutiny Panel will act to ensure that relevant lessons are taken forward. Delegates will receive feedback from the Scrutiny Panel and directly from the College on how their considerations have been acted upon. In addition, the final report will be delivered to senior police officers for them to consider, independently, how to deal with relevant lessons from the seminar. (Due June 2016.) |
| Year(s) Of Engagement Activity | 2006 |
| Description | IAFL biennial conference, Gunangzhou, China |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | 'The ellipted response felt more typical of Anthea': Performing and Detecting Online Identity Disguise Online communications pose their own unique set of challenges to investigators of paedophilia, terrorism, organised crime and other threats to the safety and security of groups and individuals. The anonymity offered by the internet has the potential to hamper policing, inevitably putting greater emphasis on forensic linguistic analysis for the tackling of online crime (Hughes et al., 2008). Preventive policing of serious crime sometimes involves deception and disguise, for example the prevention of abuse arising from online grooming of children, and peer to peer networks where abuse images of children are discussed and exchanged. Policing techniques used in these operational contexts often include assuming identities of existing community members, and of children, so that interventions and arrests can be made. This paper represents a mid-project report on a major ESRC- funded project focussing on the linguistic sources of individual and group variation in style, the description of online linguistic personas, and the disguise and detection of assumed online identities. We report on the findings of experimental tasks in which students were required to assume another's identity in an online chat context in a range of preparation conditions - with no preparation, with real-time 'over the shoulder' preparation, and with substantial revision preparation. We report on their success at assuming another's identity as gauged by the opinion of a remote 'judge' on the other end of the conversation, and on the linguistic criteria judges relied upon in making these decisions. Supplementing this work with data from real-life resolved cases of online child grooming, we provide an unparalleled insight into how particular individuals deploy and manipulate categories, and what kinds of actions they perform. The training that undercover online investigators currently receive, although broadly informed by linguistic theory, is in need of development. Thus, the implications of our work for the training provided to our police partners are drawn out and discussed. |
| Year(s) Of Engagement Activity | 2015 |
| URL | http://www.iafl.org/confarchive.php?month=July&year=2015 |
| Description | IAFL regional conference, Manila, Philippines. |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | 'There was an increase in WH- questions': accuracy, confidence and linguistic criteria for detecting impersonation in Instant Messaging: some findings from Assuming Identities Online. This paper focuses on one aspect of the two-year, ESRC-funded project Assuming Identities Online (AIO), which is now in its final stages. AIO has been concerned with the criminal context of online operations against paedophiles, and establishes what language analysis is necessary and sufficient for an undercover police officer (UCO) to successfully assume the identity of a specific victim over Instant Messaging (IM) for the purposes of drawing out and arresting a perpetrator. A series of experiments were designed within which participants were arranged into triads, and two participants were instructed to engage in conversation with each other over IM. At some point during this conversation, one of the interlocutors was substituted with the third participant. The remaining participant (the 'Judge') was asked to record their opinion on a) when the substitution had occurred; b) their confidence in this decision; and c) the criteria that led them to this decision. This paper presents the results of the judges' performances in terms of their accuracy and confidence at identifying substitution. The main focus, however, is on the linguistic criteria that they rely upon in forming these decisions. Particular attention is paid to the levels of linguistic analysis to which these comments relate, and what this tells us in terms of what facets of linguistic identity disguise are most important for UCOs to master. Our analyses have been implemented in a software assistance tool for assisting UCOs in the identity assumption tasks, which will also be demonstrated as part of this paper. |
| Year(s) Of Engagement Activity | 2016 |
| URL | http://graduateschool.ust.edu.ph/wp-content/uploads/2016/02/IAFL-Revised-Prospectus_MAR3.pdf |
| Description | IAFL regional conference, Sfax, Tunisia |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | Assuming Identities Online: Disguise and Detection of the Linguistic Persona The issue of identity and influence within transnational online communities has become a significant security concern. Online communications pose their own unique set of challenges to investigators of paedophilia, terrorism, organised crime and other threats to the safety and security of groups and individuals. Online communications provide routine means by which individuals are able to influence and radicalise, and to disseminate subversive and sometimes harmful ideologies and ways of life. The anonymity offered by the internet has the potential to hamper policing, inevitably putting greater emphasis on forensic linguistic analysis for the tackling of online crime (Hughes et al., 2008). We report on the initial stages of a collaborative project investigating the linguistic sources of individual and group variation in style, and the relationship between linguistic style and online identity performance. We report on the results of a series of experimental tasks designed to assess the level of linguistic analysis that is necessary and sufficient to describe an online linguistic persona, to the extent that it could be successfully assumed by another individual. The tasks involve engaging laypersons and postgraduate students of linguistics in live (non-criminal) online chat, and asking participants to judge whether and at which point in a chat session an interlocutor has been substituted for another with the same username. We discuss the level of accuracy and confidence that individuals can detect this substitution, and the linguistic bases for these decisions. |
| Year(s) Of Engagement Activity | 2014 |
| URL | http://10times.com/iafl-sfax |
| Description | Linguistic Identity Disguise for Online Operations |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Input on Forensic Linguistics and linguistic aspects of online identity disguise was delivered via Skype lectures to two cohorts of the Royal Canadian Mounted Police. This was followed by a simulation of a grooming operation, which in turn was followed up with Skype seminar discussions. |
| Year(s) Of Engagement Activity | 2014,2015 |
| Description | National Online Undercover Working Group Annual CPD Conference |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | Professor Tim Grant delivered a one- hour talk to participants in this event on 3rd March 2016, on the topic of "Assuming Identities Online". This covered some of the results of our experimental work, the outcomes of our analyses of training data, and a short demonstration of the IDentik software tool. |
| Year(s) Of Engagement Activity | 2016 |
| Description | Online Grooming and Preventive Justice |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Other audiences |
| Results and Impact | Tom Sorell presented his paper 'Online Grooming and Preventive Justice' at the ESRC Integrator workshop, 'Citizens Online'. The paper was also live-streamed on Twitter via Periscope to a peak audience of around 200. |
| Year(s) Of Engagement Activity | 2016 |
| URL | http://www2.warwick.ac.uk/fac/soc/pais/research/researchcentres/ierg/events/citizens_online_schedule... |
| Description | Paper on pedophile hunting and scambaiting |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Public/other audiences |
| Results and Impact | Tom Sorell presented a paper on pedophile hunting and scambaiting at a conference on digilantism at University of Montreal on 3 November 2017 |
| Year(s) Of Engagement Activity | 2017 |
| Description | Pilgrim training for online undercover officers |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | On the basis of the ongoing AIO project we have developed police training on the national online undercover course known as Pilgrim. Our input is linguistic training on online identity disguise and directly uses outputs from this project including the IDentik software. |
| Year(s) Of Engagement Activity | 2014,2015,2016 |
| Description | Policing the Virtual World |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | 14th April 2016 Tom Sorell and Chris Nathan presented papers at Canberra's Centre for Applied Philosophy and Public Ethics. They spoke alongside senior Australian police and policy officers as well as other academics. Tom Sorell presented on 'On-Line Grooming and Preventive Justice'. Chris Nathan presented on 'Proportionate Intrusions'. |
| Year(s) Of Engagement Activity | 2016 |
| URL | http://www2.warwick.ac.uk/fac/soc/pais/research/researchcentres/ierg/news/?newsItem=094d434554810e25... |
| Description | Presentation at " Human Futures and Cyber Futures", RUSI |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Tom Sorell participated in a panel alongside Richard Aldrich (Warwick) and Thomas Rid (KCL) entitled "Strategy, Social Science and Cyber in 2025" at this Warwick Intelligent Futures event. |
| Year(s) Of Engagement Activity | 2015 |
| URL | http://www2.warwick.ac.uk/fac/soc/pais/research/researchcentres/irs/wif/event/london/programme/progr... |
| Description | Presentation to serving officers of the New York Police Department |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | IERG team presentation on ethics of use of voice recognition software to about 80 serving officers of the New York Police Department at John Jay College of Criminal Justice in New York. |
| Year(s) Of Engagement Activity | 2017 |
| URL | https://warwick.ac.uk/fac/soc/pais/research/researchcentres/ierg/news/presentation_at_americal/ |
| Description | Sociolinguistics Symposium 21, Murcia |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | Spot the shibboleth(s): sociolinguistics applied to online undercover policing The anonymity afforded to individuals by the internet and online communications poses significant challenges for the policing of a wide range of criminal activities. One area of undercover online police work in which sociolinguists have been able to offer their expertise is in assisting police officers in the assumption of alternative online identities in order to apprehend offenders. Perhaps unsurprisingly, there has been a tendency for undercover operatives to rely heavily on linguistic stereotypes, both about users of Instant Messaging and about the medium of Instant Messaging itself (see Androutsopoulos, 2006), when they undertake impersonation tasks as part of investigations. Through prolonged involvement in the training of these operatives, we as sociolinguists have taken some steps to address these challenges. This paper reports on one aspect of a wider research project concerned with language and identities in the context of the policing of online grooming. Sociolinguistic experiments were designed to investigate identity assumption online and the detection of the process of impersonation. Using data collected from these experiments this paper aims to unpick exactly which areas of language use are most salient for users of Instant Messaging who are either attempting to impersonate another individual, or who are tasked with spotting the substitution of one interlocutor with another. With reference to Herring's (2004) levels of computer-mediated discourse analysis - structure, meaning, interaction and social behaviour - this paper explores how analysis at each level can prove to be a fruitful process for operatives preparing for impersonation tasks. Analysis with QSR NVivo - a suite of tools designed for the qualitative analysis of unstructured data - has allowed for a deeper understanding of the processes at work when impersonation is both performed and detected. The analyses have demonstrated that while there are some linguistic variables that are relatively easy to observe in the language patterns of others, and thus to emulate, equally there are certain variables within one's own language patterns that are prove difficult to shake. One outcome of the project is a semi-automated system for modelling linguistic personae in real time. A recommendation to arise from the analyses is that operatives model their own linguistic behaviour, as well as that of the target they plan to impersonate, before engaging in undercover work. Thus, the relevance for the training and operational contexts, and the means by which these findings will be fed back to the end user, will also be discussed. |
| Year(s) Of Engagement Activity | 2015 |
| URL | http://www.um.es/web/sociolinguistics-symposium21/ |
| Description | Undercover Policing Oversight Group, College of Policing |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | Yes |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | Questions put to senior police as part of oversight group chaired by Chief Constable Alex Marshall, including on the recent and critical HMIC report. The panel exists on an ongoing basis, for a planned period of 18 months. It provides consistent opportunities to illustrate the useful of ethics discourse in a high level policing context. |
| Year(s) Of Engagement Activity | 2014 |
| Description | Undercover policing workshop (Warwick) |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Other audiences |
| Results and Impact | Workshop on the ethics of undercover policing organised at Warwick, funded by the Department of Politics. Speakers were Seumas Miller (CSU/Delft), Christopher Nathan (Warwick), Ben Bowling (KCL), Antony Duff (Stirling); chair was Simon McKay (barrister). Participants included further academics and retired and serving police. |
| Year(s) Of Engagement Activity | 2015 |
| URL | https://www2.warwick.ac.uk/fac/soc/pais/research/researchcentres/ierg/news/undercover_workshop_sched... |
| Description | Workshop on the Ethics of Intelligence |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Other audiences |
| Results and Impact | Intelligence agencies are presently subject to far greater public and democratic scrutiny than they ever have in the past. Correspondingly, a literature is developing that probes the ethics of intelligence. How do acts of espionage relate to acts of war? Do states have carte blance in deploying electronic surveillance upon citizens or officials of other states? What kind of accountability is appropriate for intelligence officials? This seminar examines what ethical framework we best work within in order to answer those questions. Topics include how far we should depart from the just war tradition in the intelligence context, and the fundamental case for intelligence work. |
| Year(s) Of Engagement Activity | 2016 |