good morning everyone so I’m not police
I don’t work for the police I work with
the police and have done for a number of
years now actually work strictly
speaking for you my job on a day to day
basis is taking your needs your
requirements of the police as a victim
of a crime or as a taxpayer and turning
those into the sorts of things we want
the police to do making sure they’re
efficient and effective making sure even
if you are arrested for example on
suspicion of a crime that you’re treated
fairly and safely and that’s what this
talk is about it’s about turning your
needs and whatever capacity whatever way
you approach the police turning them
into reality I want you to imagine that
you have two superpowers the first
superpower is the ability to predict who
from in this room or in the wider
community is likely likely to commit a
crime the second superpower is a set
number a finite number of tokens tokens
that enable you to make interventions in
the lives of people in order to try and
prevent those crimes from happening the
tokens don’t always work but you only
have a set number with those two
superpowers in mind what decisions would
you make who would you pick what crimes
would you seek to prevent would you
choose burglary robbery sexual assaults
or terrorism or organized crime how
would you make your decision whose
advice would you seek would you see the
rise of police officers or victims of
crime or the public more broadly or
academia or research would you state
publicly that you are making these
decisions how would you expect to be
held to account for the choices you made
by picking one form of criminality to
prevent over another and here’s the one
that will really baked your noodle do
you think there are any circumstances in
which it would be appropriate for you
based on your prediction to require
somebody
except the intervention that you can use
your token for would it be possible to
compel somebody to it to redo to deal
with their potential criminality even
though they hadn’t done it yet while he
was thinking about that i’ll tell you
about somebody i’ve never met but I feel
I know quite well the man called Craig a
few years ago Craig stabbed another man
he’d never met on Broad Street his
victim nearly died he suffered what we
call in the public sector somewhat
euphemistically life-changing injuries
Craig went to jail for what he’d done I
want initial response in many ways is to
think good I’m glad Craig went to jail I
had a miserable time when he was there
for doing what he did but there’s more
to his story than you might expect Craig
was 15 when he stabbed that man on Broad
Street if you want to know more about
Craig story I’m surprisingly going to
refer you to the West Midlands Police
strategic assessment for this year
normally I wouldn’t do that because
they’re not very interesting documents
as a ruler just a list of crimes and
what we and the force plan to do about
them but this year’s is different
because it is literally Craig’s
biography from before he was born right
up to almost the present day long after
the attack took place if you go to their
website and read it well I advise you
first to kind of brace yourself because
it’s a tale of horror it’s a tale of
abuse and violence it’s a tale of
neglect you read Craig story and the
things that happen to him and you think
who would I be if that had been my
childhood you also look at the story and
you think while is there this succession
of ineffectual and largely uncoordinated
efforts by the public sector to try and
Kraig in a sense is West Midlands Police
his answer to the questions I posed to
you about superpowers it’s Craig someone
we should try and his life we should try
and change Craig story tells you two
things about the public sector though
generalizations I know but worth saying
first of all we’re not very good at
predicting who is at the greatest risk
and who poses the greatest risk and
there’s two real reasons for that first
of all we don’t exploit the data that we
hold we hold lots of data about almost
everybody and don’t properly analyze it
also we don’t share it if you have the
misfortune to read domestic homicide
reviews or serious case reviews about
the the death of children in care the
story is almost always the same the
mental health nurse didn’t know what the
GP knew the social worker didn’t know
what the police officer knew so we don’t
use our data very well to to predict
we’re also very poor at prevention
generally speaking and part of the
reason for that is that we allocate so
much resource to reaction to what
happens afterwards if you look at the
public sector the big-spending parts are
hospitals and prisons and courts and
investigations not the preventative
activity the bits where they spend the
money or the bits where we’re responding
to failure well that might be about to
change some of you have already heard of
something called predictive policing
it’s pretty straightforward actually in
its first generation iteration it is
about using crime data to predict the
places where crime will happen this is
already taking place now if you go and
talk to colleagues at british transport
police they will tell you about how
their patrol strategies use predictive
policing to make sure that they’re
police officers are in the right place
at the right time to take away the
opportunity for crimes to happen so you
will see british transport police
officers in railway stations very
steadfastly sitting and standing in
squares not much bigger than this stage
for set periods of time at certain times
a day and what that does is it takes
away the opportunity because most crime
is opportunistic it takes away the
opportunity
for the crime of robbery or shoplifting
or violent crime perhaps even to take
place that’s first-generation predictive
policing it’s here already it is already
however generating significant
controversy if you go to the American
Civil Liberties Union website there’s a
statement there by them and 15 other
human rights bodies saying how dangerous
this technology is they point out that
it could be used for even more draconian
enforcement rather than actual
prevention it could be used as a way to
catch more people incarcerate more
people rather than prevent and even more
tellingly they make the point which is
one we’ve got to think about throughout
this whole debate that the data sets on
which these predictions are being made
are built on the institutional biases of
the organizations that collected the
data so you put data in that’s biased
against particular ethnicities or
particular social groups and you come
out with predictions that are biased
against those groups too then these
concerns are very real but the
technology is racing forward we’re
moving away from prediction based solely
on place to predictions based on
individual behavior in the coming years
we will see the public sector and
policing in particular start to create
enormous fused multi-channel data lakes
with data held to common standards that
allow the application of increasingly
sophisticated algorithms that enable
genuine predictions to be made with a
fair degree of accuracy about the
behavior of individuals including their
propensity towards criminality and their
propensity towards victimhood we’re
going to see things like and they’re
starting to see the prototypes of these
now software based on real people that
predicts when they will predict the
commit their next crime where they will
be when they do so and who they’ll be
associating with real-life simulations
of real people the same technology will
also be used for developing search
patterns for repeat missing children by
knowing their associations and their
patterns of behavior will be able to
generate
tighter search patterns to make it
easier to find them your local police
officers will be equipped with targeted
intervention and prevention tools the
essentially databases of everybody on
their patch who’s likely to be a victim
and what crime will there be a victim of
who’s likely to be a criminal or commit
a criminal act and what will be the
crime that they predict that they are
predicted to it to commit our Public
Health colleagues are going to use the
same technology even now we’re starting
to see databases that enable community
health practitioners to walk down the
street and know who in which houses is
likely to die in the next two years on a
more strategic level we will see
analysis of adverse childhood
experiences as an indicator as a measure
of lightly propensity to either
victimhood or criminality and these data
sets and these predictions will become
the norm in different parts of the
public sector in policing the
opportunity presented is almost leaving
officers like giddy with excitement
because it offers the real potential to
start to address those concerns in the
public sector I talked about before of
shifting resources away from reaction
and into prevention it is transformative
I know it’s an overworked word so I’ve
actually come up with a new word for it
this word prehabilitation because it’s
about prediction prevention and
habilitation not after the event earlier
targeted intervention it’s no
coincidence that the current Chief
Constable of west midlands police in his
commencement address when he took up the
role he spoke about the duty of the
police to offer the hand of friendship
to those in need he wasn’t just talking
about victims of crime what he was
talking about is those who are at risk
who are vulnerable to becoming involved
in criminality it’s an extraordinary
opportunity but there are so many
questions to be answered and like those
that I posed at the start about what
would you do with your superpowers how
would you be held to account so that is
racing forward right now it’s happening
right now the issue then is how do we
govern it how do we run this how do we
manage this and it’s being developed in
the context of something called
operational independence if a crime is
committed the police have all
me to go away and figure out who to
investigate on how to investigate we use
the same methodology in medicine as well
whereby if you are sick if you fallen
ill you work with your GP or your
medical practitioner to figure out what
treatment you receive and no one else
has a right to tell you how to do that
no one else has the right to tell the
police how to investigate a crime that
reactive model is about dealing with
things that have been happened in the
past but we’re talking about predicting
and preventing things that are happening
in the future and my argument today is
if we’re talking about the future then
how we govern the use of that technology
must require more of us to be involved
it must require democratic legitimacy
must require a degree of public
accountability so there are a number of
models we can use to how we do that but
I see a future where politicians perhaps
standing for elected office will have to
say in their manifestos which data sets
do I trust what risks am I trying to
manage what is it that I’m trying to
prevent what are the circumstances in
which i think compulsion to deal with a
prediction is appropriate some of the
things we can learn from other ways of
doing this the police use intrusive
surveillance all the time against
against potential criminality but there
are there is a graded schedule of
authorizations that are required before
those that surveillance can take place
going all the way up to the Home
Secretary interestingly but what there
also is is an external to external
inspectorate’s whose job it is to make
sure got looking in from outside that
the way that those intrusive
surveillance technologies are used is
proportionate and lawful I think we need
we’re going to need an inspector it for
prehabilitation some kind of external
body that looks in on how these
technologies are being used if we’re
talking about compliance requiring
compliance with an intervention based on
a prediction maybe we’ll need some kind
of quasi judicial process that allows
but arguments for and against the
intervention to be considered in a
balanced and structured way it’ll give a
whole new meaning to the word pretrial
about whether or not an intervention is
justified and also there is a room for
something close to a Medical Ethics
Committee in medical research if a trial
is to go ahead or oh the introduction of
a new a new drug or a new protocol
medical ethics committees made up of lay
members of the public researchers
practitioners come together to weigh the
pros and cons and proportionality of the
intervention to see if it’s safe to see
if it’s the right thing to do all three
of those different ways of working have
something to offer in the
prehabilitation space if we call it that
but they don’t go far enough because
they are in effect unaccountable and
generally speaking they take place away
secluded from the public gaze what we
need is some is this if this technology
is going to be accepted by everybody in
this room and beyond we need much more
public legitimacy before it can blink
them implemented now I’m gonna have to
declare an interest because I think
police and crime commissioners are
important in this debate police and
crime commissioners are the local
governing body for policing I work for a
police and crime Commissioner I run a PC
sees office but they are I hope one of
the routes one of the mechanisms by
which this technology can be brought in
in a way that carries the public with us
they have a number of key advantages
first of all they’re really new police
and crime commissioners are only four
years old this year the elections first
elections were in 2012 newness is good
because newness means you can innovate
it means you’re not bound by a rule book
you’re writing the rule but we’re making
it up as we go along about what it is to
be a police and crime Commissioner the
second thing they do is they control the
money so they set the budgets for
policing and they set the local
preceptor parts of council tax that goes
to policing controlling the money’s
really important because problem money
means you can start to shift resources
away from reaction and towards
preventive services moving move
resources away from response teams and
investigation teams and towards better
support for people with mental health
needs better support for new parents who
are struggling better support for people
with substance abuse mentoring schemes
things like that PCCs can help that
shift
takes place and thirdly the obvious
point but important one is they’re
elected they voted in by the public and
that means two things first of all
they’ve got a degree of legitimacy to
make decisions they have the right to
hold the roll but even more importantly
being elected makes you responsive to
the public it means you have to listen
to the public it means you see the
public as your constituency and the
public are critical to this debate
prehabilitation preventatives policing
whatever you want to call it it’s the
best opportunity to get real value from
the public sector and bring real change
in the public sector that I’ve seen in
15 years as a public servant I’ve seen
loads of different things tried to make
the public sector more effective and
this is the one that I’m hanging more
hope than anything I’ve seen in the past
it dot offers a huge and genuine
opportunity to transform people’s lives
to actually start to get real value out
of that public spend by being able to
target it so much more effectively
against the people who need it the most
but the people who need it the most
that’s a value Laden judgment you have
to decide what you value the most in
order to decide who needs something the
most and so if those decisions about
where those interventions go ahead take
place within the context of the old
operational independence paradigm I fear
that we’re on a route towards occluded
secluded hidden decision-making where
they won’t be legitimacy where there
won’t be public accountability and the
risk is that down that road lies scandal
and failure and this huge opportunity
that we’ve got will be lost there has to
be democratically elected leadership
public publicly accountable transparent
decision-making so that we go into this
new world of using data to inform public
sector decisions in an open accountable
and genuinely transformative a
life-saving way thank you very much indeed
I don’t work for the police I work with
the police and have done for a number of
years now actually work strictly
speaking for you my job on a day to day
basis is taking your needs your
requirements of the police as a victim
of a crime or as a taxpayer and turning
those into the sorts of things we want
the police to do making sure they’re
efficient and effective making sure even
if you are arrested for example on
suspicion of a crime that you’re treated
fairly and safely and that’s what this
talk is about it’s about turning your
needs and whatever capacity whatever way
you approach the police turning them
into reality I want you to imagine that
you have two superpowers the first
superpower is the ability to predict who
from in this room or in the wider
community is likely likely to commit a
crime the second superpower is a set
number a finite number of tokens tokens
that enable you to make interventions in
the lives of people in order to try and
prevent those crimes from happening the
tokens don’t always work but you only
have a set number with those two
superpowers in mind what decisions would
you make who would you pick what crimes
would you seek to prevent would you
choose burglary robbery sexual assaults
or terrorism or organized crime how
would you make your decision whose
advice would you seek would you see the
rise of police officers or victims of
crime or the public more broadly or
academia or research would you state
publicly that you are making these
decisions how would you expect to be
held to account for the choices you made
by picking one form of criminality to
prevent over another and here’s the one
that will really baked your noodle do
you think there are any circumstances in
which it would be appropriate for you
based on your prediction to require
somebody
except the intervention that you can use
your token for would it be possible to
compel somebody to it to redo to deal
with their potential criminality even
though they hadn’t done it yet while he
was thinking about that i’ll tell you
about somebody i’ve never met but I feel
I know quite well the man called Craig a
few years ago Craig stabbed another man
he’d never met on Broad Street his
victim nearly died he suffered what we
call in the public sector somewhat
euphemistically life-changing injuries
Craig went to jail for what he’d done I
want initial response in many ways is to
think good I’m glad Craig went to jail I
had a miserable time when he was there
for doing what he did but there’s more
to his story than you might expect Craig
was 15 when he stabbed that man on Broad
Street if you want to know more about
Craig story I’m surprisingly going to
refer you to the West Midlands Police
strategic assessment for this year
normally I wouldn’t do that because
they’re not very interesting documents
as a ruler just a list of crimes and
what we and the force plan to do about
them but this year’s is different
because it is literally Craig’s
biography from before he was born right
up to almost the present day long after
the attack took place if you go to their
website and read it well I advise you
first to kind of brace yourself because
it’s a tale of horror it’s a tale of
abuse and violence it’s a tale of
neglect you read Craig story and the
things that happen to him and you think
who would I be if that had been my
childhood you also look at the story and
you think while is there this succession
of ineffectual and largely uncoordinated
efforts by the public sector to try and
Kraig in a sense is West Midlands Police
his answer to the questions I posed to
you about superpowers it’s Craig someone
we should try and his life we should try
and change Craig story tells you two
things about the public sector though
generalizations I know but worth saying
first of all we’re not very good at
predicting who is at the greatest risk
and who poses the greatest risk and
there’s two real reasons for that first
of all we don’t exploit the data that we
hold we hold lots of data about almost
everybody and don’t properly analyze it
also we don’t share it if you have the
misfortune to read domestic homicide
reviews or serious case reviews about
the the death of children in care the
story is almost always the same the
mental health nurse didn’t know what the
GP knew the social worker didn’t know
what the police officer knew so we don’t
use our data very well to to predict
we’re also very poor at prevention
generally speaking and part of the
reason for that is that we allocate so
much resource to reaction to what
happens afterwards if you look at the
public sector the big-spending parts are
hospitals and prisons and courts and
investigations not the preventative
activity the bits where they spend the
money or the bits where we’re responding
to failure well that might be about to
change some of you have already heard of
something called predictive policing
it’s pretty straightforward actually in
its first generation iteration it is
about using crime data to predict the
places where crime will happen this is
already taking place now if you go and
talk to colleagues at british transport
police they will tell you about how
their patrol strategies use predictive
policing to make sure that they’re
police officers are in the right place
at the right time to take away the
opportunity for crimes to happen so you
will see british transport police
officers in railway stations very
steadfastly sitting and standing in
squares not much bigger than this stage
for set periods of time at certain times
a day and what that does is it takes
away the opportunity because most crime
is opportunistic it takes away the
opportunity
for the crime of robbery or shoplifting
or violent crime perhaps even to take
place that’s first-generation predictive
policing it’s here already it is already
however generating significant
controversy if you go to the American
Civil Liberties Union website there’s a
statement there by them and 15 other
human rights bodies saying how dangerous
this technology is they point out that
it could be used for even more draconian
enforcement rather than actual
prevention it could be used as a way to
catch more people incarcerate more
people rather than prevent and even more
tellingly they make the point which is
one we’ve got to think about throughout
this whole debate that the data sets on
which these predictions are being made
are built on the institutional biases of
the organizations that collected the
data so you put data in that’s biased
against particular ethnicities or
particular social groups and you come
out with predictions that are biased
against those groups too then these
concerns are very real but the
technology is racing forward we’re
moving away from prediction based solely
on place to predictions based on
individual behavior in the coming years
we will see the public sector and
policing in particular start to create
enormous fused multi-channel data lakes
with data held to common standards that
allow the application of increasingly
sophisticated algorithms that enable
genuine predictions to be made with a
fair degree of accuracy about the
behavior of individuals including their
propensity towards criminality and their
propensity towards victimhood we’re
going to see things like and they’re
starting to see the prototypes of these
now software based on real people that
predicts when they will predict the
commit their next crime where they will
be when they do so and who they’ll be
associating with real-life simulations
of real people the same technology will
also be used for developing search
patterns for repeat missing children by
knowing their associations and their
patterns of behavior will be able to
generate
tighter search patterns to make it
easier to find them your local police
officers will be equipped with targeted
intervention and prevention tools the
essentially databases of everybody on
their patch who’s likely to be a victim
and what crime will there be a victim of
who’s likely to be a criminal or commit
a criminal act and what will be the
crime that they predict that they are
predicted to it to commit our Public
Health colleagues are going to use the
same technology even now we’re starting
to see databases that enable community
health practitioners to walk down the
street and know who in which houses is
likely to die in the next two years on a
more strategic level we will see
analysis of adverse childhood
experiences as an indicator as a measure
of lightly propensity to either
victimhood or criminality and these data
sets and these predictions will become
the norm in different parts of the
public sector in policing the
opportunity presented is almost leaving
officers like giddy with excitement
because it offers the real potential to
start to address those concerns in the
public sector I talked about before of
shifting resources away from reaction
and into prevention it is transformative
I know it’s an overworked word so I’ve
actually come up with a new word for it
this word prehabilitation because it’s
about prediction prevention and
habilitation not after the event earlier
targeted intervention it’s no
coincidence that the current Chief
Constable of west midlands police in his
commencement address when he took up the
role he spoke about the duty of the
police to offer the hand of friendship
to those in need he wasn’t just talking
about victims of crime what he was
talking about is those who are at risk
who are vulnerable to becoming involved
in criminality it’s an extraordinary
opportunity but there are so many
questions to be answered and like those
that I posed at the start about what
would you do with your superpowers how
would you be held to account so that is
racing forward right now it’s happening
right now the issue then is how do we
govern it how do we run this how do we
manage this and it’s being developed in
the context of something called
operational independence if a crime is
committed the police have all
me to go away and figure out who to
investigate on how to investigate we use
the same methodology in medicine as well
whereby if you are sick if you fallen
ill you work with your GP or your
medical practitioner to figure out what
treatment you receive and no one else
has a right to tell you how to do that
no one else has the right to tell the
police how to investigate a crime that
reactive model is about dealing with
things that have been happened in the
past but we’re talking about predicting
and preventing things that are happening
in the future and my argument today is
if we’re talking about the future then
how we govern the use of that technology
must require more of us to be involved
it must require democratic legitimacy
must require a degree of public
accountability so there are a number of
models we can use to how we do that but
I see a future where politicians perhaps
standing for elected office will have to
say in their manifestos which data sets
do I trust what risks am I trying to
manage what is it that I’m trying to
prevent what are the circumstances in
which i think compulsion to deal with a
prediction is appropriate some of the
things we can learn from other ways of
doing this the police use intrusive
surveillance all the time against
against potential criminality but there
are there is a graded schedule of
authorizations that are required before
those that surveillance can take place
going all the way up to the Home
Secretary interestingly but what there
also is is an external to external
inspectorate’s whose job it is to make
sure got looking in from outside that
the way that those intrusive
surveillance technologies are used is
proportionate and lawful I think we need
we’re going to need an inspector it for
prehabilitation some kind of external
body that looks in on how these
technologies are being used if we’re
talking about compliance requiring
compliance with an intervention based on
a prediction maybe we’ll need some kind
of quasi judicial process that allows
but arguments for and against the
intervention to be considered in a
balanced and structured way it’ll give a
whole new meaning to the word pretrial
about whether or not an intervention is
justified and also there is a room for
something close to a Medical Ethics
Committee in medical research if a trial
is to go ahead or oh the introduction of
a new a new drug or a new protocol
medical ethics committees made up of lay
members of the public researchers
practitioners come together to weigh the
pros and cons and proportionality of the
intervention to see if it’s safe to see
if it’s the right thing to do all three
of those different ways of working have
something to offer in the
prehabilitation space if we call it that
but they don’t go far enough because
they are in effect unaccountable and
generally speaking they take place away
secluded from the public gaze what we
need is some is this if this technology
is going to be accepted by everybody in
this room and beyond we need much more
public legitimacy before it can blink
them implemented now I’m gonna have to
declare an interest because I think
police and crime commissioners are
important in this debate police and
crime commissioners are the local
governing body for policing I work for a
police and crime Commissioner I run a PC
sees office but they are I hope one of
the routes one of the mechanisms by
which this technology can be brought in
in a way that carries the public with us
they have a number of key advantages
first of all they’re really new police
and crime commissioners are only four
years old this year the elections first
elections were in 2012 newness is good
because newness means you can innovate
it means you’re not bound by a rule book
you’re writing the rule but we’re making
it up as we go along about what it is to
be a police and crime Commissioner the
second thing they do is they control the
money so they set the budgets for
policing and they set the local
preceptor parts of council tax that goes
to policing controlling the money’s
really important because problem money
means you can start to shift resources
away from reaction and towards
preventive services moving move
resources away from response teams and
investigation teams and towards better
support for people with mental health
needs better support for new parents who
are struggling better support for people
with substance abuse mentoring schemes
things like that PCCs can help that
shift
takes place and thirdly the obvious
point but important one is they’re
elected they voted in by the public and
that means two things first of all
they’ve got a degree of legitimacy to
make decisions they have the right to
hold the roll but even more importantly
being elected makes you responsive to
the public it means you have to listen
to the public it means you see the
public as your constituency and the
public are critical to this debate
prehabilitation preventatives policing
whatever you want to call it it’s the
best opportunity to get real value from
the public sector and bring real change
in the public sector that I’ve seen in
15 years as a public servant I’ve seen
loads of different things tried to make
the public sector more effective and
this is the one that I’m hanging more
hope than anything I’ve seen in the past
it dot offers a huge and genuine
opportunity to transform people’s lives
to actually start to get real value out
of that public spend by being able to
target it so much more effectively
against the people who need it the most
but the people who need it the most
that’s a value Laden judgment you have
to decide what you value the most in
order to decide who needs something the
most and so if those decisions about
where those interventions go ahead take
place within the context of the old
operational independence paradigm I fear
that we’re on a route towards occluded
secluded hidden decision-making where
they won’t be legitimacy where there
won’t be public accountability and the
risk is that down that road lies scandal
and failure and this huge opportunity
that we’ve got will be lost there has to
be democratically elected leadership
public publicly accountable transparent
decision-making so that we go into this
new world of using data to inform public
sector decisions in an open accountable
and genuinely transformative a
life-saving way thank you very much indeed
Please follow and like us: