Q&A
Plans for biometric data and predictive models in MyCity
1:41:38
·
4 min
Council Member Gutiérrez inquires about plans to incorporate biometric data and predictive models into the MyCity portal. Commissioner Fraser discusses the current stance on biometric data and the potential use of predictive models.
- Fraser states that incorporating biometric data is not currently planned for MyCity
- He explains that any multifactor authentication would be secured by platforms and not housed by the city
- Regarding predictive models, Fraser confirms that they are considering their use for improving service delivery and targeting areas of need
- He emphasizes that predictive capabilities would be based on need and demand, not individual behavior
- Fraser assures that role-based access control would be implemented to restrict data access, including for agencies like the NYPD
- The discussion highlights the balance between using data for service improvement and protecting individual privacy
Jennifer Gutiérrez
1:41:38
Can you share if the administration has any plans to incorporate biometric data or identification technologies into the MyCiti portal?
Matthew Fraser
1:41:46
So at this moment, incorporating biometric data is not on the forefront of our approach.
1:41:54
When even when you look at things that use multifactor authentication, those those methods are secured by platforms.
1:42:01
It's not that's not data that's housed by the city.
1:42:03
So at this moment, there's no current plan or no there's no current plan or projected plan to incorporate biometric data.
Jennifer Gutiérrez
1:42:12
Great.
1:42:14
Is there any plan to include predictive models to the MyCiti portal?
1:42:21
I know the mayor This is early back in 2022.
1:42:26
Wanted to kind of combine all agency metrics similar to to ComStat.
1:42:33
So curious if there is a similar intention for my city?
Matthew Fraser
1:42:38
Yeah.
1:42:38
So we we we would want to keep agencies not just accountable for the work that's been completed, but the quality of work that's been done.
1:42:50
So predictive models, analytics, to a large scale is all part of what's being done in this world.
1:42:56
If you look at this is like what we're well, if you look at the scope of what we're doing in terms of rendering benefits to a new degree, you wanna ensure that you have the maximum amount of efficacy in deploying those benefits.
1:43:08
And also you're reaching the populations of people that need that.
1:43:11
In order to assess whether we're doing a good job or not, we have to be able to look back.
1:43:17
And in order to assess how much we can grow, we need the capability to look forward as well.
1:43:21
So we expect that in both of those spaces to have those capabilities in place.
Jennifer Gutiérrez
1:43:27
Sorry.
1:43:27
And would it be to predict behavior or to determine eligibility?
1:43:33
Is there a sense that can you get to have specificity on, like, what the
Matthew Fraser
1:43:37
So so behavior, when I say predictive capabilities is predictive, but capabilities based on need and demand.
1:43:46
So those that are applying, are they actually being serviced?
1:43:50
Those with demand?
1:43:51
Are they actually adult years where there are demand?
1:43:55
How effectively are we targeting those areas?
Jennifer Gutiérrez
1:43:58
Okay.
1:44:00
I just I know that in other cities they've attempted to incorporate, like, predictive models into benefits and services, and there's just there's been negative side effects.
1:44:09
So just wanting to kind of hear from you the steps the steps you take, just like profiling.
1:44:20
Right?
1:44:20
Sharing some of that predictive behavior data with PD, for example, just wanting to, like, hear it from you kind of, like, a a strong yes or no or maybe if that is, like, a a potential feature of the portal.
Matthew Fraser
1:44:33
Yeah.
1:44:33
I I I'd say for sure, it's a potential feature of the the portal.
1:44:36
We would need the capability to in order to effectively serve, not just deal with situations as they arise, but predict to a degree how much is likely gonna be necessary.
1:44:48
A good example for that is if you know with the pipeline for early childhood, what childcare looks like and you wanted to protect the amount of pre k c's that you would need during any given period, having the capability to use the information that you have to say, based on what we can see the total volume of kids are.
1:45:04
This is this is potentially what the need would be for this particular thing.
Jennifer Gutiérrez
1:45:07
And would that access be available to every single agency including PD?
1:45:12
No.
Matthew Fraser
1:45:13
No.
1:45:13
So in situations like this, this is where role based access control comes from and attribute based access control.
1:45:20
What that basically means is that you may work in DSS and you may need access to a specific portion of a record, right, a person's record, but you don't need to see everything in that record.
1:45:31
Right?
1:45:31
Based on your role and based on the attribute in that record, making sure that you only have access to the things that you need to make your determination and not any more than that.
Jennifer Gutiérrez
1:45:41
Okay.
1:45:43
Specifically, would would that also apply to the police department?
Matthew Fraser
1:45:49
I'm Would
Jennifer Gutiérrez
1:45:50
would the role based access?
Matthew Fraser
1:45:52
Oh, for sure.
1:45:53
For sure.
1:45:53
No no no one is exempt from that process.
