Digital Policing: The

Ethical Issues Arising

from Digital Policing
(Revised Version November 2021)

1
 Foreword
My generation was the first in history to be apprenticed to the younger generation
coming up behind. The first but not the last. I was taught how to use a computer and
a smartphone by my children – something that amused and annoyed them in equal
measure. Why was I so slow in grasping the rudiments of the new technologies! Why
wasn’t I teaching them? It defied the natural order. I say this simply to indicate the
pace of change in the digital world that we are experiencing and the challenges it is
constantly bringing.

All of which opens up new possibilities for policing even as it opens up new
opportunities for criminals. It also presents us with new ethical dilemmas and
choices.

I am fortunate in having an Independent Ethics Panel (IEP) to advise me, and South
Yorkshire police, on ethical issues. So I asked them to consider the range of new
technologies that are presenting themselves, the kinds of ethical issues they raise
and, crucially, the sorts of questions we should be asking about them. I wanted
them to look at this from the perspective of the public whom I represent and who are
likely to have views on the use by the police of at least some new technologies.

In the future, we may still have a bobby on the beat, but she will be carrying a
mobile device and using data that previous generations of officers would never
have had access to. There will be other officers and staff, less visible, using data in
many new ways to make policing more effective and efficient. And there will be new
technologies with the potential to make communication between police and public
better. All needs to be done ethically, and we need to know what questions we
should ask.

This is what the IEP were tasked to think about and this is their report. I thank them
for it.

Dr Alan Billings

Police and Crime Commissioner for South Yorkshire

2 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

Digital Policing
Re-reading our report, first drafted several years ago, it is striking how rapidly things
are changing.

In particular Artificial Intelligence (AI) is evolving at a pace and in directions which

raise fresh concerns. It is no longer simply a question of how best to apply the
emerging technologies but increasingly of who or what is making the decisions.
There is growing uncertainty about the true capabilities of what is becoming
available.

In addition, security, not only of data but also of design, manufacture and
procurement, has gained prominence.

This is the context in which the ethical questions raised in the Independent Ethics
Panel report are both urgent and important. The new technologies offer genuine
potential to effective and efficient policing, but the challenge is to deploy them in
a way which maintains the trust and confidence of the public whilst enhancing the
service provided by the police.

As we move further into the Digital Age, the ethical dimension of policing must
remain at the top of the agenda. This report is our first response to Dr Billings’
request. We hope we have identified the right questions and that the answers will be
the subject of constant scrutiny.

Michael Lewis

Chair

Independent Ethics Panel

3
 Digital policing: the ethical
issues arising from digital
policing
(Revised Version November 2021)

devices and resources that generate, store or

Remit process data. Well known examples include social
At the Independent Ethics Panel meeting on media, online games, multimedia and mobile
27 June 2019, members agreed to undertake phones (Department of Education 2019).
a piece of work which would look at the ethical
issues which may be raised when considering the The internet and related digital technology has led
introduction of new technology to support digital to vast increases in the amount of data available
policing. The full agreed terms of reference can be about individuals and it is growing exponentially
found at appendix 2. each year.

The aim was to set up a review which would Internet users generate about 2.5 quintillion bytes
culminate in the creation of a set of ethical of data each day and it comes from a range of
principles and a checklist of questions focussing sources. Social media users spent the equivalent
on the ethical issues to be considered when of 2.8 million years online in 2018, generating 33%
introducing any form of digital policing technology. of the data.
It was agreed that the review would:
A Law Commission Consultation on Search
■ L
 ook at the different types of technologies Warrants (2018) reported that individuals now
that were understood by people and have more information about themselves on their
organisations under the umbrella of digital electronic devises than is in their homes.
policing.
A House of Commons Science and Technology
■ C
 onsider the ethical issues raised by these Committee report on Biometrics and Forensics
technologies. (2018) stated that existing databases of facial
■ T
 ake account of the ethical concerns across images such as driving licences, passports and
different groups. police custody records cover 90% of the adult
■ H
 elp inform South Yorkshire Police (SYP) population.
Policy and decision making around digital
Data analytics has been developed to exploit these
policing.
data sets and big business and agencies like the
Police are being urged to engage with these new
Background technologies.
Digital technologies are electronic tools, systems,

4 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

Digital Policing
The Policing Vision 20251 sets out the future of Machine Learning (ML): Machine learning is a
policing up to 2025. On the subject of digital particular type of AI. The House of Lords Select
policing, the Vision says: Committee on Artificial Intelligence (2018) defined
it as follows: “Machine Learning is a field of study
“Digital policing will: that gives computers the ability to learn without
being explicitly programmed. It utilises a range of
1. Make it easier for the public to make contact with
algorithms to produce models from data sets than
the police wherever they are in the country;
can then be used to provide predictive analytics.”
2. Enable us to make better use of digital
Algorithms: These are series of mathematical
intelligence and evidence and;
instructions for programming computers to perform
3. T
 ransfer all material in a digital format to the calculations, solve problems, or produce models
criminal justice system.” than can be used to make predictions. Algorithms
are being used to help predict an outcome,
The Policing Vision statement includes a range of map crime, identify particular individuals or
ways in which technology can be used in policing, create a risk score for individuals. It includes
from facilitating and recording interactions with the those technologies that might employ machine
public, to transferring and sharing data externally learning where, over time and based on the data
as well as more specialist applications some of being collected, a computer adjusts and creates its
which are outlined in figure 1 on page 6. own algorithms.

These approaches are often categorised as either, Examples of some of these technologies that
“place-based predictive policing” or “person- have been or are being used in policing in the UK
based predictive targeting”. Our focus is on include:
identifying ethical issues related to the use of such
technologies and primarily on those which currently ■ D
 urham Constabulary’s Harm Assessment
appear to be contentious with sections of the Risk Tool (HART) to help assess the risk of
public,(or have the potential to be so), namely: future offending by those arrested.
■ K
 ent Police’s use of a “Gangs Matrix” to
1. Artificial Intelligence technologies that use identify and keep under observation those at
algorithms in predictive policing and risk of committing gang-related violence and
potential victims of gang-related violence.
2. C
 ameras and automated Facial
Recognition Technology. ■ T
 rials of live automated facial recognition
systems by the Metropolitan and South
The same basic ethical principles however, should Wales police forces, whereby a “watch list”
apply more widely when considering other types of people are mapped against live scans of
of technology that could come under the digital faces in public places and crowds to identify
policing umbrella, including those not yet invented those wanted by the police.
or even contemplated. ■ A
 von and Somerset Police’s use of the Qlik
system to assess risk and predict crime.

Definitions ■ W
 est Midlands Police’s work on a “National
Analytics Solution” aimed at combining a
Artificial Intelligence (AI): While there is yet no range of data from across partnerships to
commonly agreed definition of AI pioneer Marvin prevent harm, reduce criminality and protect
Minsky’s definition is widely used and adopted the vulnerable.
here; “Artificial Intelligence is the science of making
computers do things that require intelligence when
done by humans.”2 This definition raises ethical
questions about the principles associated with AI,
namely, that when humans are involved in decision
making and the use of intelligence, they can be
held accountable for the judgement, accuracy and
reliability in their decision making. There is less
clarity about how to hold computers to account,
yet accountability is crucial if errors are made.

1
developed by the Association of Police and Crime Commissioners and the National Police Chiefs’ Council
2
Understanding Artificial Intelligence Ethics and Safety: A Guide for the Responsible Design and Implementation of AI systems in the Public Sector. The Alan
Turing Institute

5
 The following figure reports on a Deloitte Survey of nine forces to collect their views on the relevance of
digital technologies and what they are using or may invest in.

Figure 1: Police Organisations are experimenting widely with new technologies

Tom Gash and Richard Hobbs (2019) Deloitte Survey. Policing in a digital world

The benefits and concerns

around digital policing
technologies
the data sets being used in these technologies,
Police perspective for example concerns about data protection and
The use of artificial intelligence as part of digital security and privacy issues. These are coming from
policing can be of real benefit to the police and the MPs, the Information Commissioner’s Office and
public. Being able to assess risk quickly, identify the Surveillance Commissioner, as well as specific
wanted suspects or missing persons efficiently and campaigning groups such as Liberty, Big Brother
identify and predict potential “hot-spot” areas of Watch and the media.
particular types of crime can help the police operate
more efficiently and keep the public safe. This is There are also practical delivery challenges in coping
particularly true against a backdrop of changing with the quantities of digital data available and also the
demand and crime patterns. adequacy of the existing police IT systems to deal with
them and to interact with each other. While this may
Against these benefits, however, there are concerns not immediately seem like an ethical issue, provision
about the use of artificial intelligence in policing and of reliable tools for the workforce to utilise so that they

6 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

can work effectively and safely is crucial. A survey of in an open and transparent way. This should help
18 forces by Virgin Media in 2018 reported that only 2 South Yorkshire Police in its decision making around
forces were confident in this area. the trialling or operational deployment of digital
technologies so that either or both can be achieved
Currently, the police use the National Decision without eroding public trust in policing.
Making Model (NDMM) which has at its heart the
Code of Ethics to guide their decision-making.
The model systematically uses the knowledge, Resources used in the review:
skills and experience of officers to problem solve.
 iterature searches (See references in appendix
■ L
Utilising analytics could be perceived by officers as
1);
disregarding their knowledge and experience. Careful
thought needs to be given about how and when ■ R
 eviewed legislation and outcomes from recent
digitally generated solutions are accepted or when legal cases;
they are rejected in favour of human decision-making. ■ E
 xamined government, national association
There is little research on how acceptable these bodies and police force strategies, policies
predictive techniques are to officers being asked to and guidance, including the National Decision
use them. One German study evaluating the use of Making Model (NDMM) and the College of
predictive policing for residential burglaries found that Policing Code of Ethics;
opinions about usefulness were almost equally split ■ S
 ought input from South Yorkshire Police’s
(Gerstner, September, 2018). (SYP) Internal Ethics Committee on emerging
findings prior to completing the final report;
Public concerns ■ A
 ttended a workshop on ethics in digital policing
and contributed to development of work
The Independent Digital Ethics Panel for Policing nationally in this area;
(IDEPP)3 has expressed concerns about the
■ E
 ngaged with relevant stakeholders including
inadequacy of the current national oversight
academics at Sheffield University developing
mechanisms for the use of such technology in
ethical oversight of artificial intelligence in
policing. They are also concerned about the,
policing and the Department of Digital, Culture,
“commercial and political pressures to adopt
Media and Sport’s Centre for Data Ethics and
algorithmic techniques for crime prevention and
Innovation.
reduction which need to be counterbalanced by
clear articulation of requirements, circumspect
The information collected has been gathered into
procurement, ethical scrutiny, and informed debate,
themes, to enable identification of some key ethical
particularly on the question of bias” (Independent
principles that need to be considered.
Digital Ethics Panel for Policing, 2018). If these issues
are not recognised and addressed, there is a danger
that public scepticism about these digital technologies The Digital Policing Ethical
will erode trust and confidence in the police where
these technologies are trialled or introduced Principles and Ethical
operationally.
Questions
Concerns about privacy, human rights and issues This work has identified one fundamental starting
around data bias, data quality and data protection are question to be addressed, namely, “Is the technology
the issues raised most often in the public arena. There being deployed in the public interest?” The
are other public concerns too such as: assumption is that this should always be the case.
Six key digital policing ethical principles are identified
■ Keeping data up to date. which then form the basis for developing a set of
■ E
 nsuring data is used appropriately in line with questions to test the ethics of a particular digital
the General Data Protection Rules (GDPR). policing technology. These six principles identified are:
■ M
 aking sure data is stored and retained
correctly. ■ Legality
■ Accuracy
Knowledge of these concerns helps to set the context ■ Openness and Transparency
within which ethical questions about use are raised. ■ Embedded in Good Governance
A failure to consider these concerns in an open and
transparent way can damage public trust in digital ■ Fairness
policing technologies. ■ Accountability

This review is thus timely, in identifying ethical Underpinning these is the concept of proportionality
principles and highlighting the ethical questions which can and should be considered when policing

3
 he formation of IDEPP was encouraged by the National Police Chief’s Council but is an independent panel. It is made up of experts drawn from a range
T
of disciplines including serving law enforcement professionals, academics, civil society thinkers and other legal/policy professionals. The Panel provides
insight and challenge to a range of areas associated with digital policing.

7
 decisions are being made. Proportionality also features a number of times within the subsequent list of ethical
questions.

At the national level police forces have implemented the College of Policing’s Code of Ethics (CoE) and the
following table maps this code against the digital policing ethical principles.

Fit with College of Policing Code of Ethics (CoE)

Digital Policing College of Policing
Ethical Principle Code of Ethics Principle
Legality The legal issues around digital technology implementation can be very complex and expert
legal consideration is sometimes necessary. Application of the law must be carried out through
the lens of ethical principles.
Accuracy Objectivity: You make choices on evidence and your best professional judgement
Fairness: You treat people fairly
Respect: You treat everyone with respect
Openness and Openness: You are open and transparent in your actions and decisions
transparency Honesty: You are truthful and trustworthy
Embedded in good Leadership: You lead by good example
governance Openness: You are open and transparent in your actions and decisions
Honesty: You are truthful and trustworthy
Objectivity: You make choices on evidence and your best professional judgement
Integrity: You always do the right thing
Selflessness: You act in the public interest
Fairness Fairness: You treat people fairly
Respect: You treat everyone with respect
Integrity: You always do the right thing
Accountability Accountability: You are answerable for your decisions actions and omissions

The Ethical Questions

The ethical questions that have been developed against the digital policing principles are provided in the
following table to form the basis of an ethical specification that can be used by those considering the trial
or deployment of digital policing technologies.

In developing these questions, it has been recognised that there is relevant guidance and legislation
already in place that may help decision makers, but this has been developed incrementally and may not
always focus overtly on ethical considerations. What these questions aim to do is to start to draw some of
the legislation and guidance together into an ethical focus around digital policing technologies.

The principles and related questions developed as part of this work will be kept under review as national
legislation and guidance around digital policing technologies develop further and as such the following
questions will be reviewed and refined in order to stay relevant and up to date.

8 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

Ethical Digital Ethical Questions Considerations and Resources
Policing
Principle
Legality 1. Is the technology being used in the public interest? • Protection of Freedom Act 2012
2. Is there a transparent legal basis to justify the technology use? • Regulation of Regulatory Powers
3. Is there an agreed policing purpose which justifies the use of the technology? Act 2000

4. Is the data being used limited to the agreed policing purpose? • Data Protection Act 2018

5. Is there the potential to interfere with the privacy of individuals and how will this be minimised? • Common law

6. Is the use of the data proportionate to the risk of harm? • Human Rights Act

7. Is there a record of how the technology complies with relevant national guidance? • Freedom of Information Act

8. Has the use of the technology been assessed and recorded against the principles of • Equality Act 2010

• Necessity • The governing principles of the

“Biometrics and Forensic Group”
• Proportionality
• Surveillance Camera Code
• Data minimisation
• Home Office Biometrics Strategy
• Data protection and
• Forensic Science Regulators
• Data retention
Guidance
9. Has “the duty to give reasons” been taken into account for technologies that indicate decisions, for
• Forensic Science Regulator’s Code
example on the management of offenders?
of Practice
10. Is there a clear, publicised process in place for those affected by decisions made by digital technology to
• The forces definition of “Wanted by
challenge those decisions
the Police”
11. How will decisions be made about whether consent from individuals is necessary?
12. If necessary how will consent be obtained from individuals affected by the technology as opposed to
assumed consent?
13. Are/will there be records in place to show that the retention of data is proportionate?
14. How will it be assured that human decision makers are not blindly following algorithmic results and so
unlawfully “fettering their discretion”?

Accuracy 1. Does the specification for the technology match the policing need? • SYP Procurement Policy
2. Are there regular system and data audits built into work plans so that algorithms and data are checked • SYP Workforce Strategy
regularly for accuracy to avoid bias?
• College of Policing guidance on the
3. Is there an expert provided by the supplier who can give details of algorithms operating in the digital
Management of Police Information
technology?
(MOPI)
4. A re officers trained to understand, deploy and interpret algorithmic tools in combination with their
professional expertise?
5. A re you able to define the risks, consequences and acceptability of inaccurate forecasts or matches and
mitigate accordingly?
6. Is there a system with clear and unambiguous criteria for compiling and recording data sets and watch
lists? This is especially important where joint data sets are used or data sets from more than one
organisation are merged.
7. Are there arrangements in place to ensure the data used in technologies is of good quality?
8. Are there checks in place to ensure the quality of 3rd party data as well as checks that 3rd party data is/
has been collected for the same use as the intended policing need?
9. Are there mechanisms in place to ensure that the data the technology relies on is current and regularly
updated and complies with data retention policies?
10. Is there a training plan in place to make sure those operating the technology have the relevant skills,
expertise and knowledge to ensure:
• data integrity
• good data quality
• accurate interpretation of results
• an awareness of algorithmic aversion
• an awareness of potential injustices that may be caused by inappropriate decisions/responses by the
technology?

9
 Ethical Digital Ethical Questions Considerations and Resources
Policing
Principle
Openness and 1. A re arrangements in place to communicate to the public, the rationale for the use of the digital
transparency policing technology (both in trials and operationally) to ensure trust and confidence is built and
maintained?

2. Are engagement channels defined, publicised and available for public/stakeholder feedback?

3. A re there plans to trial and evaluate new technologies using public and stakeholder feedback
before, large scale roll-out?

4. H
 ow will information be made available about the decision making rules of any algorithm, scoring
or outcome used in the technology?

5. How will algorithms and algorithmic systems allow for maximum control and amendment by SYP?

6. W
 hat process is/will be in place so that those affected by algorithmic decisions are able to
challenge those decisions (scoring/rating/judgement) etc?

7. What are the minimum standards that will be set for transparency and intelligibility of algorithms
and decision making of the technology?

8. Is there an expert provided by the supplier who can provide details of the algorithms operating in
the digital technology?

9. Is there a clear basis for the retention of data?

Embedded in 1. A re there governance and oversight arrangements in place for both trials and operational Governance arrangements should
good governance deployment of technologies including a senior responsible named officer? ensure that trials and deployments of
digital policing technology :
2. How are governance arrangements currently enhancing and improving those algorithmic
• are legal
technologies already in use?
• are necessary
3. Where the technology is jointly undertaken and operated with partners and third parties, are there • are proportionate
governance arrangements in place to ensure clear protocols and clearly assigned responsibilities • are safe and secure
particularly around data management? • a chieve benefits balanced against
potential intrusiveness
4. W
 here technologies are procured from a third party, are there robust contract management
• a lways Include up to date datasets
arrangements in place to monitor and manage performance?
and algorithms
5. D
 o contracts allow SYP the right to access, use and amend source code and data if necessary? • a re carried out by officers with the
necessary skills, knowledge and
expertise in that particular technology
and any relevant legal knowledge.

Fairness 1. Are there mechanisms in place to ensure that the algorithm is operating fairly and identifies potential • Centre For Data Ethics and
algorithmic bias (towards/against certain groups)? This will include ensuring that the data used in Innovation - interim report –review
algorithms or watch lists are impartial and do not inadvertently import unacceptable gender and racial into bias in algorithmic decision
bias. making (policing strand)
2. Will an equality impact assessment be undertaken before the trial or implementation of the technology to
https://www.gov.uk/government/
consider its impact on all communities?
publications/interim-reports-
3. Are there mechanisms to review regularly the wider social, political and cultural implications of digital from-the-centre-for-data-
technology as well its policing effectiveness and impact on staff? ethics-and-innovation/
4. W
 hat arrangements have been made to inform data subjects as to whether their data will be retained? interim-report-review-into-bias-in-
5. H
 ow will the fairness checks be communicated to the public? algorithmic-decision-making

Accountability 1. How and in what circumstances will human decision makers be able to override algorithmic decision Consider use of Turing Institute’s
making to improve the objectivity of decision making? ethical platform for responsible AI
2. How will human decision makers gain the skills and knowledge to be aware of and address: project delivery.

• The accurate interpretation of results

• The potential for algorithmic deference/aversion by human decision makers
• The potential injustices that may be caused by inappropriate decisions/responses by the
technology
• The need to ensure there is no fettering of discretion?

10 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

Final Reflections
The close match between the Digital Policing Ethical Principles and the College of Policing’s Code of Ethics could facilitate its
operational use within the Police National Decision Making Model when new digital technology projects are being considered.
The ethical questions can be matched to each area thus ensuring that at each stage ethical questions related to the technology
are being addressed.

Police National Decision Making Model

Gather information and

intelligence

Assess threat and risk

Take action and review
and develop a working
what happened
strategy

CODE OF
ETHICS

Identify options and Consider powers and

contingencies policey

• Stage 1 – Gather information about the problem in hand. • Stage 4 – Armed with all of the information regarding the
Not only should you work out what you do know but what problem and any policies and other legislations that may
you do not know. You will use this information, gathered in exist, you are in a position to draw up a list of options. You
stage 1, throughout the rest of the process and once your should also use this opportunity to develop a contingency
decisions are being assessed and judged after the event. plan or a series of contingencies that can provide you with a
backup plan if things don’t get exactly to plan.
• Stage 2 – Determine the threat, its nature, and its extent
so that you can assess the situation and make the right • Stage 5 – Once you have determined the most appropriate
decisions. Do you need to take the necessary action straight action, it is time to put this in place. Perform the most
away or is this an ongoing problem? What is the most likely desirable action and, if necessary, begin the process again
outcome and what would be the implications? Are the police to get the best results possible. Review the process and
the right people to deal with the problem and are you best determine whether or not you could have done things better
equipped to help resolve the problem at hand or would and what you would do in the future if you were faced with a
somebody else be better? similar, or the same problem.

• Stage 3 – Knowing what the problem is, you will need to

determine what powers you and the police have in order
to combat the problem. Which powers will be needed and
will the required powers and policies need any additional
or specialist manpower to be instigated and introduced? Is
there any legislation that covers the process.

11
 Appendix 1
Documents Reviewed 2019). Accessed at https://www.wired.co.uk/
article/police-facial-recognition-south-wales-
Alan Turing Institute (2018). Ethics advisory report court-decision
for West Midlands Police
House of Commons Science and Technology
Babuta, A., Oswald, ., & Rinik, C. (2018). Whitehall Committee (2016) Robotics and Artificial
Report 3-18: Machine Leaning Algorithms and Intelligence report. https://www.parliament.
Police Decision-Making. Accessed from: https:// uk/business/committees/committees-a-z/
rusi.org/sites/default/files/201809_whr_3-18_ commons-select/science-and-technology-
machine_learning_algorithms.pdf.pdf committee/inquiries/parliament-2015/robotics-
and-artificial-intelligence-inquiry-15-16/
Biometrics and Forensics Ethics Group (Feb 2019). publications/
Police use of live facial recognition technology:
ethical issues. https://www.gov.uk/government/ House of Lords Select Committee on Artificial
publications/police-use-of-live-facial- Intelligence (2018). AI in the UK: ready, willing and
recognition-technology-ethical-issues able? https://publications.parliament.uk/pa/
ld201719/ldselect/ldai/100/100.pdf
Civica (2019) ( Previously Election Reform Services).
Response to Centre for Data Ethics and Innovation Information Commissioner’s Office (2017). Big Data,
(part of Dept. for Digital, Culture, Media and Sport) Artificial Intelligence, Machine Learning and Data
consultation on algorithmic bias in policing and Protection.
justice https://www.gov.uk/government/news/
investigation-launched-into-potential-for-bias- Kearns, I., & Muir, R. (2919). The Police Foundation:
in-algorithmic-decision-making-in-society Data-Driven Policing and Public Value. http://www.
police-foundation.org.uk/project/data-driven-
College of Policing: Police National Decision policing-and-public-value-new/
Making Model
Leslie, D. (2019) Understanding Artificial Intelligence
College of Policing: Code of Ethics Ethics and Safety: A guide for the responsible
design and implementation of AI systems in the
Couchman, H., ( Jan ,2019). Predictive Policing and public sector” The Alan Turing Institute https://
the Threat to Our Rights. London: Liberty. www.turing.ac.uk/sites/default/files/2019-06/
understanding_artificial_intelligence_ethics_
Department for Digital, Culture, Media & Sport. and_safety.pdf
(2017). Unlocking the power of data in the UK
economy and improving public confidence in its Oswald, M. Algorithm assisted decision making in
use. (2017 the public sector. Philosophical Transactions of the
Royal Society Accepted: 25 June 2018 https://cris.
Digitial Policing Summit (2019) - Highlighting the winchester.ac.uk/ws/files/2032839/272985_
need to address ethical issues re data use. https:// Oswald_AlgorithmicPublicSector_
www.gov.uk/government/publications/uk- withstatement.pdf
digital-strategy/7-data-unlocking-the-power-
of-data-in-the-uk-economy-and-improving- Oswald, M., Grace, J., Urwin, S. & Barnes, G.C.
public-confidence-in-its-use (2018). Algorithmic risk assessment policing
models: lessons from the Durham HART model
Fussey, P. & Murray, D. (July 2019). Independent and ‘Experimental’ proportionality, Information &
Report on the London Metropolitan Police Communications Technology Law, 27:2, 223-250,
Service’s Trial of Live Facial Recognition DOI: 10.1080/13600834.2018.1458455
Technology: The Police use of Automated Facial
Recognition Technology with Surveillance Camera Organisation for Economic Co-operation and
Systems. (Economic Research Council, University Development (OECD) ( May 2019). Human-centred
of Essex and Human Rights Centre) principles on artificial intelligence https://www.
oecd.org/going-digital/ai/principles/
Gerstner, D. (2018). Predictive Policing in the
Context of Residential Burglary: An Empirical https://www.cityforum.co.uk/wp-content/
Illustration on the Basis of a Pilot Project in Baden- uploads/2019/11/On-the-day-Agenda-Fifth-
Württemberg, Germany. European Journal of Digital-Policing-Summit.pdf
Security Researches, 3, 115–138. doi:10.1007/
s41125-018-0033-0 Policing by machine (Feb 2019): Authorr Hannah
Couchman. Published by Liberty. https://www.
High Court Judgement Summary on the use of libertyhumanrights.org.uk/policy/report-
facial recognition by South Wales Police (Sept policing-machine

12 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

Public trust and data-driven policing – Other Research:
Police Professional (2019). https://www.
policeprofessional.com/industry-viewpoint/ Workshop: Exploring Ethical Standards in The Use
public-trust-and-data-driven-policing/ of Data Analytics in Policing – Sheffield Hallam
University
Rex, H. (Dec. 2016) How digital tech will change
how we fight crime. Public Sector. https://www. Civica (2019) ( Previously Election Reform Services).
itproportal.com/features/how-digital-tech-will-
change-how-we-fight-crime/ House of Lords Select Committee on Artificial
Intelligence (2018). ”AI in the UK: ready, willing and
Saunders, J., Hunt, P., Hollywood, J. S. (2016). able?”
Predictions put into practice: a quasi-experimental
evaluation of Chicago’s predictive policing pilot. Independent Digital Ethics Panel for Policing (2018):
Journal of Experimental Criminology, 12(3):347–371. Response to Centre for Data Ethics and Innovation
(part of Dept. for Digital, Culture, Media and Sport)
Strom, K., for US Department of Justice (2017). consultation on algorithmic bias in policing and
Research on the Impact of Technology on Policing justice
Strategy in the 21st Century.
Public trust and data-driven policing –
http://www.ncjrs.gov/App/publications/ Police Professional (2019). https://www.
abstract.aspx?ID=273320 policeprofessional.com/industry-viewpoint/
public-trust-and-data-driven-policing/
The Law Society ( June 2019). Algorithm Use in The
Criminal Justice System (Summary) https://www. Rex, H. How digital tech will change how we fight
lawsociety.org.uk/support-services/research- crime, December 2016, Public Sector.
trends/algorithm-use-in-the-criminal-justice-
system-report/ Strom, K., for US Department of Justice (2017).
Research on the Impact of Technology on Policing
The Policing Vision 2025 developed by the Strategy in the 21st Century
Association of Police and Crime Commissioners
and the National Police Chief’s Council The House of Lords Select Committee on Artificial
Intelligence
Virgin Media (2018). Police Survey of IT Systems.
https://www.virginmediabusiness.co.uk/ The Policing Vision 2025 developed by the
insights/digital-policing-report/ Association of Police and Crime Commissioners
and the National Police Chief’s Council

ICO Annual Track Survey, April 2016

Robotics and Artificial Intelligence report, October

2016, House of Commons Science and Technology
Committee

Department for Digital, Culture, Media & Sport

Unlocking the power of data in the UK economy
and improving public confidence in its use.
Published 1 March 2017

Saunders J, Hunt P, Hollywood JS (2016)

Predictions put into practice: a quasi-experimental
evaluation of Chicago’s predictive policing pilot. J
Exp Criminol 12(3):347–371

The Police Foundation: Data-Driven Policing and

Public Value Ian Kearns and Rick Muir, March
2019.

Digital Policing Summit - Highlighting the need to

address ethical issues re data use in particular

Predictive Policing in the Context of Residential

Burglary: An Empirical Illustration on the Basis of a
Pilot Project in Baden-Württemberg, Germany. Eur
J Secur Res 3, 115–138 (2018) doi:10.1007/s41125-
018-0033-0 Gerstner D

13
 Appendix 2
The Vision is not prescriptive in what technologies could come
Digital Policing: The Ethical Issues under the umbrella of Digital Policing but from the description
Arising from Digital Policing of what Digital Policing will do, the sort of technologies
involved could be wide ranging.
Terms of Reference The Independent Ethics Panel want to ensure that specific
ethical issues, pertaining to a range of technologies, are
1. Overall aim: identified, considered and appropriately addressed by SYP
when technologies supporting Digital Policing are being
To support South Yorkshire Police (SYP) in understanding considered for introduction to support the Policing Vision
the range of ethical questions that may be raised when 2025.
considering introducing new digital technology to support
Digital Policing and provide assurance to the public that ethical
issues have been considered and addressed. 3. In Scope
In doing so, the review should: The review will:

■ Identify ethical issues raised around Digital Policing and ■ L

 ook at the different types of technologies that are
ensure they are considered, as appropriate, by SYP. understood by people and organisations that come
under the umbrella of Digital Policing. As this is in a
■ C
 reate an ethical checklist for police officers and staff state of continuous development, the aim is therefore to
to use as part of any project introducing Digital Policing identify points which may be relevant for future users as
techniques. well as present ones. (See also point 5 below).
■ C
 onsider the ethical issues raised by these different
2. Background: types of technologies.
■ T
 ake account of the ethical concerns across different
The Policing Vision 20254 sets out the future of policing up to groups of residents.
2025. On the subject of Digital Policing, the Vision says:
■ H
 elp inform SYP policy and decision making around
“Digital policing will: digital policing.

■ m
 ake it easier for the public to make contact with the
police wherever they are in the country; 4. Out of Scope
■ e
 nable us to make better use of digital intelligence and ■ Non ethical issues will not be considered.
evidence and;
■ Deciding final policy on digital policing.
■ transfer all material in a digital format to the criminal
■ Implementation of digital policing systems unless the
justice system.”
manner or method of implementation raises ethical
issues.
The Vision does not explicitly link Digital Policing to particular
ethical issues, however the Vision says generally on ethics:

■ T
 he values of the police service are detailed within the 5. Methodology and Evidence Gathering
Code of Ethics which set out the nine principles which Having acknowledged earlier that the range of technologies
underpin and strengthen the existing procedures and under the “digital policing” umbrella could be wide ranging and
regulations for ensuring standards of professional in continuous development (see point 3 above), it will be useful
behaviour for both police officers and police staff. as a starting point to take initially:
■ R
 espect for human rights will be central to everything we
do, as is the commitment to policing by consent. ■ Automated Facial Technology and;
■ T
 he use of evidence based practice and the Code of ■ A
 rtificial Intelligence technologies that use algorithms in
Ethics will be embedded and inform day to day policing predictive policing as technologies to consider ethical
practice. considerations against.
■ P
 olicing must embed legitimacy, trust and confidence,
underpinned by the Code of Ethics. Once defined, these considerations can be tested against
other types of technology that could come under the “digital
policing” umbrella.

4
developed by the Association of Police and Crime Commissioners and the National Police Chief’s Council

14 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING

Evidence gathering will include: 7. Timescales
■ Relevant literature and study review. 5 ■ June 2019 – scope to be agreed at IEP
■ L
 egislative changes and outcomes from ■ Complete work during Summer 2019
decided legal cases.
■ D
 iscussion of review outcomes with SYP
■ R
 eview of relevant government, national during September 2019
association bodies and police force
■ Report to IEP October/November 2019
strategies/policies.
■ Input from SYP’s Internal Ethics Panel on
emerging findings prior to final reporting. 8. Resources:
■ Independent Ethics Panel Chair
6. Outputs ■ Lead Panel Member
■ A
 short report reviewing the ethical issues ■ SYP Ethics Lead
around digital policing. ■ O
 PCC Support Officer: Evaluation and
■ A
 checklist of ethical issues to consider Scrutiny Officer
that can be used flexibly and assist SYP in ■ OPCC Sponsor: Head of Governance
thinking of ethics when introducing any form
of Digital Policing technology.
■ A
 checklist available on the OPCC’s public
website to inform the public and partners of
ethical considerations to be considered.

5
Resources already available to review:
• The Law Society of England & Wales investigation published 4 June 2019
• London Police Ethics Panel’s report(s) on Live facial recognition technology
• Surveillance Camera Commissioner: Police Use of AFR, published March 2019
• West Midlands PCC’s Ethics Committee Minutes 3 April 2019: Integrated Offender Management Model etc.
• Work by Marion Oswald of the University of Winchester
• Ongoing study by DCMS’ Centre for Data Ethics & Innovation, especially an enquiry announced on 20 March 2019
• Liberty: Policing by machine, published 4 Feb 2019

15
 facebook.com/SouthYorkshirePoliceandCrimeCommissioner

@sypcc

@sypcc

SYPCC Media
4091-06/23

16 DIGITAL POLICING: THE ETHICAL ISSUES ARISING FROM DIGITAL POLICING