An invaluable tool to prevent retail crime or an invasion of privacy that can be used to track you without your consent?
Justice Minister Paul Goldsmith wants to allow the use of facial recognition technology in the fight against retail crime, despite acknowledging 'tensions' with people’s right to privacy.
Retail NZ CEO Carolyn Young explains how this technology gathers data to prevent retail crime - as well as the risks that come with this method.
LISTEN ABOVE
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:09):
You're listening to a podcast from News Talk ZEDB. Follow
this and our wide range of podcasts now on iHeartRadio.
Speaker 2 (00:16):
Justice Minister Paul Goldsmith wants to allow the use of
facial recognition technology in the fight against retail crime, despite
acknowledging tensions with people's right to privacy. To discuss this,
it's great to have Retail n Z Chief executive Carolyn
Young on the phone. Carolyn, very good afternoon to you.
Speaker 3 (00:34):
Carol, A good afternoon.
Speaker 4 (00:35):
So how does this trial facial recognition technology work in
a store?
Speaker 3 (00:42):
Well, it's I'll do my best to describe it and
a quick couplement. It's as quite complex system. So basically
to want to put a facial recognition trial in your store,
you need to have some reason why you would do that.
So that would be because you've got a number of
people that are continuing to cause abuse and violence and
(01:02):
aggression in store. The primary reason it's been put on
faces for staff safety. A lot of violent and aggressive
behavior we see in retail these days, and so what
stores have what food Stuffs have done, and they've established
in a number of stores people that have offended in store,
whether it's that they have been aggressive and violent towards staff,
(01:26):
or potentially they may have stolen and been aggressive and
those sorts of behaviors. They've trespassed them, they've ben offended.
They've got a photo of them from the truspass. Notice.
They load that photo and the photo is a headshot only,
so it doesn't include anything below your chin, so no shoulders,
no body, no nothing, It's only your head. And so
(01:47):
the headshot gets loaded into the facial resgnition technology system.
And in the stores that they were random stores that
were selected to do the trial as part of the
requirements from the Privacy Commissioner, And in those twenty five
stores that did the trial, they had large signs out
the front of the store at every entrance telling people
(02:10):
that they had that their fatal recognition was in play
in the store, and that if they didn't want to
come in store, that was their obviously their ability to
not come in, but they were notified before they entered
store so that they had the choice about whether they
came in the store. When you walk in the store,
and you or I walk in, we have never offended
in that particular store, our image is deleted immediately. There's
(02:34):
no storing of any data if, on the other hand,
John Smith, who last time he was in the store
he spat and kicked that star. Yeah, he has loaded
up into the system. He comes back into store after
he's been trespassed, an alert goes to a device that
(02:56):
tells the store that John Smith has just entered. Now
there's a couple of things that happened here. Firstly, the
store can traffic lighter a person that they enter in,
so they might actually like someone to say, really violent
and aggressive, do not approach, call the police immediately. You know,
based on the previous actions that this person has had
(03:18):
in that store, and so you would know. And part
of the way you create a safer work environment is
that you would then know that you've got an extremely
violent and aggressive person in store, and he's not to
be he or she is not to be approached. You
might have somebody else who is, you know, maybe they
had a bad day, they haven't offended previously, they were
(03:40):
abusive in store, but they have been trespassed for three
months or something like that, so they might well be approachable.
So you could easily approach them as they have come
to start to have come in the store and say,
excuse me, John Smith will use John Smith again, Well,
maybe you'll use free dag. Excuse me, free dag. But
you know you've been trespassed. You're not meant to be
(04:00):
in here. And you know eighty to ninety percent of
the times you approach somebody upon entry and acknowledge them around,
the actions are taken previously and they shouldn't be in stored.
They will turn around and walk out and they apologize,
and then there might be some other people that you
might just monitor in store. And what facial recognition does
(04:23):
is that it's a biometric that it takes a view
of the face that matches up your facial features. The
minimum match was set at ninety percent, and what food
Stuff's recognized is that the quality of the photo you
put into the system needs to be quite high to
ensure that you get an accurate match. So the threshold
(04:44):
was ninety percent. If anything was over ninety four or
ninety five percent, they knew that virtually every time it
was one hundred percent match between ninety and ninety four
percent has to take more care because there is an
element of risk in there, and so with every alert
that they get, they'd have two trained people would look
at that photo and consider whether or not it was
(05:06):
the person that was in their system.
Speaker 4 (05:09):
So when going going forward with this situation, you say,
this was just sort of people splitting and violent offenders
and starch. So shoplifters aren't the ones being targeted in
this trial. But would you see going forward that shoplifters
would have their faces loaded into this facial recognition technology?
Speaker 3 (05:27):
Yeah, so shoplifters may well have also been included because
people that are feeling are often violent and aggressive. What's that?
So there's a combination of things that the primary reason
was to keep the staff safe, and by understanding who's
in the store enables you to create a safer work
environment for your staff and for your customers. So it ultimately,
(05:50):
you know, does a dual purpose, right, It will make
the store safer and will reduce the amount offending because
most of the people that are being violent and abusive
and aggressive are also offending to a certain degree in store.
I mean, you can't go and say to your work calling,
you can't go and split at the nose or punch
them off. And videos of a four second interaction between
(06:13):
a customer and a staff member where the customer just
punches the staff member in the face and it's got
accelerated in four seconds. They walk up to the counter,
they have a conversation, they smack them straight away. I've
seen other images where people's gone in store and they
have set fire. They've done us in a store to
steal a trolley load of gods. So they've gone in
(06:35):
there and that puts every staff member in danger. Right,
if your business is going to be set alight, that
is quite stressful environment to be in, so you don't
want to help the host people coming back in your store,
and you want to protect yourself. So it's a dual purpose, right,
it protects The primary thing is that it protects the
staff and creates a safe environment. But at the same time,
(06:57):
these are people that are offending, so it will reduce
the level of crime, and it can be used as
a preventative tool rather than a reactive tool after the event.
And a key thing is that if you can approach
someone within the first minute of them being in the store,
they have not got themselves worked up into a frenzy
and they're not going to be as aggresive and violent
(07:18):
as you are. If you get them straight away, if
they've had time to mooch around the store and they
know that they've got to get your adrenaline up to
force through a trolley worth of goods and not pay
for it. If you're doing that before they've even got
anything in a trolley or in a bag, you get
a much better outcome and certainly the outcomes and the
trials for that time and time again, we're.
Speaker 4 (07:38):
Talking to Carolyn Young, Retail New Zealand Chief Executive. So
is it security that approaches the people? So you have
security guards on staff and they approach them. It's not
up to the checkout person to approach these people that
are ping.
Speaker 3 (07:54):
Well, no, it's not up to the checkout person. So
it'll be people in store that have been identified and
have been trained. So it could be a security guard
if there's no security guard on judy, because it's not
there's two people. So you'll probably have three or four
people across environment of a supermarket that are trained at
any point in time that could do the identification and
would get better alert So, for example, if it's at
(08:17):
a time when a security guard isn't on duty, then
it would be it'll be two other senior people within
the store, but it wouldn't be a checkout operator that
would be asked to ask someone to leave. It's someone
that's trained and skilled and doing so and can identify,
you know, as this person violent and aggressive and do
we need to call a police straight away? And that's
a really that's probably one of the greatest tools is
(08:39):
knowing that people that are really aggressive can be addressed
in a way in which you don't have to you
call the police. They will come because they know that
you've already identified someone that's been high level offending previously.
Speaker 2 (08:51):
What are the laws in place for the storage of
that data, caront because a lot of people they may
unfairly be freaking out they come to upt images of
Minority Report nineteen eighty four. But as the strict laws
and fail safes in place that that data is deleted,
as you say, if nobody does anything wrong and certainly
can't get into the wrong hands.
Speaker 3 (09:11):
Yeah, so there's a couple of things. I guess, well,
there's quite a lot of things not try and remember
to cover everything off. But firstly, for those people that
are not on the alert list, your images to meet
deleted immediately because there's nothing to match against. So within
the second of you being in the store, your image
is deleted. The person who has created the alert or
(09:33):
the image that's created the alert, that file is deleted
at midnight that night. This is how the trial was
a set up with the privacy commissioner. So that means
that the store has got till midnight to complete a
report around what happened, was it the right person? You know,
all of those sorts of things and file that. But
(09:54):
then that file is then removed from the system and
that's controlled through you know, third parties that controlled all
of that stuff, so it's not stored anywhere you can't
go into. And the computer that the the financial the
facial recognition technology sits on as a separate computer. It's
not on a network, it's not on Wi Fi, it's
(10:16):
not on any system that is part of a supermarket
network or within the store. It's completely standalone. So and
it doesn't store anything beyond midnight of an alert from
that person. So some of those things might be tweaked
a little bit. You can imagine as an example, in
a supermarket, if you are matching a face and only
(10:40):
the face you and it's a ninety six percent match,
then you know that's that person come into store, very
hard to find that person on a Saturday afternoment, you've
got hundreds of people in the store because you don't
you don't have any idea what they're wearing at all.
Did you find it because all you've got from the
image as a face? Yeah?
Speaker 4 (11:01):
Did you find because you said this was heavily sign
signed and signposted for people coming into the stores that
are on the trial. Did you find that it affected
numbers going in? Because that's really if people are comfortable,
if people are going in and they're happy to be
have their their faces recognized. That I mean round about
way of saying, didn't affect the business?
Speaker 3 (11:25):
No, not in a negative way, not like that. No,
not at all. There's always people that will question whether
they're comfortable with something like that technology being used, and
they may choose not to go install. But there was
no absolutely less well less than one percent. And then
there's always ways in which people can access further information
via a website around how it's run and what happens
(11:46):
in a faq on the website and those sorts of things.
So the key is making sure that you communicate really
well and really strongly with your customers so that they're
confident when they're coming install that they're being well looked
after and that people aren't using their images in any
way that's inappropriate.
Speaker 4 (12:02):
Well, just two questions to finish, Thank you so much
for your time. Will the data be shared with the
police via the Aurora network? And secondly, what about people
wearing masks and head coverings or disguises?
Speaker 3 (12:15):
Great question, right, Let's let's deal with the seat of
the last question first. So what you know, especially if
you've got security sitting outside your store, if someone's coming
in with a mask, hotty up and dark glasses, that's
that's screams and cat leaders travel right, telltale signs, and
so what normally happens. The social recognition system normally has
(12:38):
between two and four cameras in store, that's all right
at the entry, not linked to any of the CCTV or anything.
So if they don't capture them in the first two cameras,
often someone that has that hasn't been stopped by security
for the way they dressed with the hot and the
mask and the gloves and the shades and everything. They'll
(13:02):
take the shades off and the mask off the minute
they get in the store, and they can get picked
up on the second camera. But it is certainly a
high alert situation. If you see someone like that, it
stills trouble with the capital T.
Speaker 2 (13:12):
Gotcha and just on the AURA network, Carolyn.
Speaker 3 (13:15):
Yes, No, it's a completely separate system. You can't transfer
any data across to any other system. It's not part
of that. If somebody is ascending in store, you would
be using your CCTV images to report those through. And
remembering that not every retailer uses AURA, So whether you'd
be reporting through the one oh five or one one
one or through AURA, nothing comes off the facial Refnition
(13:41):
Technology platform to enable you to do that, you have
to use your CCTV for that.
Speaker 2 (13:45):
Gotcha, Carolyn, Thank you very much for having to chat
with us. Very interesting. That is retail in z CEO
Carolyn Young.
Speaker 1 (13:53):
For more from News Talk SEDB, listen live on air
or online, and keep our shows with you wherever you
go with our podcasts on IR Radio.
You're listening to a podcast from News Talk ZEDB. Follow
this and our wide range of podcasts now on iHeartRadio.
Speaker 2 (00:16):
Justice Minister Paul Goldsmith wants to allow the use of
facial recognition technology in the fight against retail crime, despite
acknowledging tensions with people's right to privacy. To discuss this,
it's great to have Retail n Z Chief executive Carolyn
Young on the phone. Carolyn, very good afternoon to you.
Speaker 3 (00:34):
Carol, A good afternoon.
Speaker 4 (00:35):
So how does this trial facial recognition technology work in
a store?
Speaker 3 (00:42):
Well, it's I'll do my best to describe it and
a quick couplement. It's as quite complex system. So basically
to want to put a facial recognition trial in your store,
you need to have some reason why you would do that.
So that would be because you've got a number of
people that are continuing to cause abuse and violence and
(01:02):
aggression in store. The primary reason it's been put on
faces for staff safety. A lot of violent and aggressive
behavior we see in retail these days, and so what
stores have what food Stuffs have done, and they've established
in a number of stores people that have offended in store,
whether it's that they have been aggressive and violent towards staff,
(01:26):
or potentially they may have stolen and been aggressive and
those sorts of behaviors. They've trespassed them, they've ben offended.
They've got a photo of them from the truspass. Notice.
They load that photo and the photo is a headshot only,
so it doesn't include anything below your chin, so no shoulders,
no body, no nothing, It's only your head. And so
(01:47):
the headshot gets loaded into the facial resgnition technology system.
And in the stores that they were random stores that
were selected to do the trial as part of the
requirements from the Privacy Commissioner, And in those twenty five
stores that did the trial, they had large signs out
the front of the store at every entrance telling people
(02:10):
that they had that their fatal recognition was in play
in the store, and that if they didn't want to
come in store, that was their obviously their ability to
not come in, but they were notified before they entered
store so that they had the choice about whether they
came in the store. When you walk in the store,
and you or I walk in, we have never offended
in that particular store, our image is deleted immediately. There's
(02:34):
no storing of any data if, on the other hand,
John Smith, who last time he was in the store
he spat and kicked that star. Yeah, he has loaded
up into the system. He comes back into store after
he's been trespassed, an alert goes to a device that
(02:56):
tells the store that John Smith has just entered. Now
there's a couple of things that happened here. Firstly, the
store can traffic lighter a person that they enter in,
so they might actually like someone to say, really violent
and aggressive, do not approach, call the police immediately. You know,
based on the previous actions that this person has had
(03:18):
in that store, and so you would know. And part
of the way you create a safer work environment is
that you would then know that you've got an extremely
violent and aggressive person in store, and he's not to
be he or she is not to be approached. You
might have somebody else who is, you know, maybe they
had a bad day, they haven't offended previously, they were
(03:40):
abusive in store, but they have been trespassed for three
months or something like that, so they might well be approachable.
So you could easily approach them as they have come
to start to have come in the store and say,
excuse me, John Smith will use John Smith again, Well,
maybe you'll use free dag. Excuse me, free dag. But
you know you've been trespassed. You're not meant to be
(04:00):
in here. And you know eighty to ninety percent of
the times you approach somebody upon entry and acknowledge them around,
the actions are taken previously and they shouldn't be in stored.
They will turn around and walk out and they apologize,
and then there might be some other people that you
might just monitor in store. And what facial recognition does
(04:23):
is that it's a biometric that it takes a view
of the face that matches up your facial features. The
minimum match was set at ninety percent, and what food
Stuff's recognized is that the quality of the photo you
put into the system needs to be quite high to
ensure that you get an accurate match. So the threshold
(04:44):
was ninety percent. If anything was over ninety four or
ninety five percent, they knew that virtually every time it
was one hundred percent match between ninety and ninety four
percent has to take more care because there is an
element of risk in there, and so with every alert
that they get, they'd have two trained people would look
at that photo and consider whether or not it was
(05:06):
the person that was in their system.
Speaker 4 (05:09):
So when going going forward with this situation, you say,
this was just sort of people splitting and violent offenders
and starch. So shoplifters aren't the ones being targeted in
this trial. But would you see going forward that shoplifters
would have their faces loaded into this facial recognition technology?
Speaker 3 (05:27):
Yeah, so shoplifters may well have also been included because
people that are feeling are often violent and aggressive. What's that?
So there's a combination of things that the primary reason
was to keep the staff safe, and by understanding who's
in the store enables you to create a safer work
environment for your staff and for your customers. So it ultimately,
(05:50):
you know, does a dual purpose, right, It will make
the store safer and will reduce the amount offending because
most of the people that are being violent and abusive
and aggressive are also offending to a certain degree in store.
I mean, you can't go and say to your work calling,
you can't go and split at the nose or punch
them off. And videos of a four second interaction between
(06:13):
a customer and a staff member where the customer just
punches the staff member in the face and it's got
accelerated in four seconds. They walk up to the counter,
they have a conversation, they smack them straight away. I've
seen other images where people's gone in store and they
have set fire. They've done us in a store to
steal a trolley load of gods. So they've gone in
(06:35):
there and that puts every staff member in danger. Right,
if your business is going to be set alight, that
is quite stressful environment to be in, so you don't
want to help the host people coming back in your store,
and you want to protect yourself. So it's a dual purpose, right,
it protects The primary thing is that it protects the
staff and creates a safe environment. But at the same time,
(06:57):
these are people that are offending, so it will reduce
the level of crime, and it can be used as
a preventative tool rather than a reactive tool after the event.
And a key thing is that if you can approach
someone within the first minute of them being in the store,
they have not got themselves worked up into a frenzy
and they're not going to be as aggresive and violent
(07:18):
as you are. If you get them straight away, if
they've had time to mooch around the store and they
know that they've got to get your adrenaline up to
force through a trolley worth of goods and not pay
for it. If you're doing that before they've even got
anything in a trolley or in a bag, you get
a much better outcome and certainly the outcomes and the
trials for that time and time again, we're.
Speaker 4 (07:38):
Talking to Carolyn Young, Retail New Zealand Chief Executive. So
is it security that approaches the people? So you have
security guards on staff and they approach them. It's not
up to the checkout person to approach these people that
are ping.
Speaker 3 (07:54):
Well, no, it's not up to the checkout person. So
it'll be people in store that have been identified and
have been trained. So it could be a security guard
if there's no security guard on judy, because it's not
there's two people. So you'll probably have three or four
people across environment of a supermarket that are trained at
any point in time that could do the identification and
would get better alert So, for example, if it's at
(08:17):
a time when a security guard isn't on duty, then
it would be it'll be two other senior people within
the store, but it wouldn't be a checkout operator that
would be asked to ask someone to leave. It's someone
that's trained and skilled and doing so and can identify,
you know, as this person violent and aggressive and do
we need to call a police straight away? And that's
a really that's probably one of the greatest tools is
(08:39):
knowing that people that are really aggressive can be addressed
in a way in which you don't have to you
call the police. They will come because they know that
you've already identified someone that's been high level offending previously.
Speaker 2 (08:51):
What are the laws in place for the storage of
that data, caront because a lot of people they may
unfairly be freaking out they come to upt images of
Minority Report nineteen eighty four. But as the strict laws
and fail safes in place that that data is deleted,
as you say, if nobody does anything wrong and certainly
can't get into the wrong hands.
Speaker 3 (09:11):
Yeah, so there's a couple of things. I guess, well,
there's quite a lot of things not try and remember
to cover everything off. But firstly, for those people that
are not on the alert list, your images to meet
deleted immediately because there's nothing to match against. So within
the second of you being in the store, your image
is deleted. The person who has created the alert or
(09:33):
the image that's created the alert, that file is deleted
at midnight that night. This is how the trial was
a set up with the privacy commissioner. So that means
that the store has got till midnight to complete a
report around what happened, was it the right person? You know,
all of those sorts of things and file that. But
(09:54):
then that file is then removed from the system and
that's controlled through you know, third parties that controlled all
of that stuff, so it's not stored anywhere you can't
go into. And the computer that the the financial the
facial recognition technology sits on as a separate computer. It's
not on a network, it's not on Wi Fi, it's
(10:16):
not on any system that is part of a supermarket
network or within the store. It's completely standalone. So and
it doesn't store anything beyond midnight of an alert from
that person. So some of those things might be tweaked
a little bit. You can imagine as an example, in
a supermarket, if you are matching a face and only
(10:40):
the face you and it's a ninety six percent match,
then you know that's that person come into store, very
hard to find that person on a Saturday afternoment, you've
got hundreds of people in the store because you don't
you don't have any idea what they're wearing at all.
Did you find it because all you've got from the
image as a face? Yeah?
Speaker 4 (11:01):
Did you find because you said this was heavily sign
signed and signposted for people coming into the stores that
are on the trial. Did you find that it affected
numbers going in? Because that's really if people are comfortable,
if people are going in and they're happy to be
have their their faces recognized. That I mean round about
way of saying, didn't affect the business?
Speaker 3 (11:25):
No, not in a negative way, not like that. No,
not at all. There's always people that will question whether
they're comfortable with something like that technology being used, and
they may choose not to go install. But there was
no absolutely less well less than one percent. And then
there's always ways in which people can access further information
via a website around how it's run and what happens
(11:46):
in a faq on the website and those sorts of things.
So the key is making sure that you communicate really
well and really strongly with your customers so that they're
confident when they're coming install that they're being well looked
after and that people aren't using their images in any
way that's inappropriate.
Speaker 4 (12:02):
Well, just two questions to finish, Thank you so much
for your time. Will the data be shared with the
police via the Aurora network? And secondly, what about people
wearing masks and head coverings or disguises?
Speaker 3 (12:15):
Great question, right, Let's let's deal with the seat of
the last question first. So what you know, especially if
you've got security sitting outside your store, if someone's coming
in with a mask, hotty up and dark glasses, that's
that's screams and cat leaders travel right, telltale signs, and
so what normally happens. The social recognition system normally has
(12:38):
between two and four cameras in store, that's all right
at the entry, not linked to any of the CCTV or anything.
So if they don't capture them in the first two cameras,
often someone that has that hasn't been stopped by security
for the way they dressed with the hot and the
mask and the gloves and the shades and everything. They'll
(13:02):
take the shades off and the mask off the minute
they get in the store, and they can get picked
up on the second camera. But it is certainly a
high alert situation. If you see someone like that, it
stills trouble with the capital T.
Speaker 2 (13:12):
Gotcha and just on the AURA network, Carolyn.
Speaker 3 (13:15):
Yes, No, it's a completely separate system. You can't transfer
any data across to any other system. It's not part
of that. If somebody is ascending in store, you would
be using your CCTV images to report those through. And
remembering that not every retailer uses AURA, So whether you'd
be reporting through the one oh five or one one
one or through AURA, nothing comes off the facial Refnition
(13:41):
Technology platform to enable you to do that, you have
to use your CCTV for that.
Speaker 2 (13:45):
Gotcha, Carolyn, Thank you very much for having to chat
with us. Very interesting. That is retail in z CEO
Carolyn Young.
Speaker 1 (13:53):
For more from News Talk SEDB, listen live on air
or online, and keep our shows with you wherever you
go with our podcasts on IR Radio.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com