Q&A
Responsibility for AI risk assessment and potential public disclosure
1:05:58
·
173 sec
Council Member Gutiérrez inquires about the responsibility for AI risk assessments and the possibility of making the results public. Alex Foard from OTI explains the current state of the process and the considerations involved.
- OTI is committed to building out the risk assessment process
- Agencies will play a key role as business owners driving the need and solution development
- It's too early to determine the exact step-by-step process for risk assessment
- Transparency is a principle, but it's unclear what specific information will be available for public disclosure
- The process aims to balance transparency with agency-specific obligations and constraints
Jennifer Gutiérrez
1:05:58
And do you all have a sense of who who's, like, responsible for the risk assessment?
1:06:05
Is it OTI?
1:06:06
Is it the agency using the system?
1:06:10
Where are you all, I guess, in in that piece?
Alex Foard
1:06:13
Sure.
1:06:13
So it's probably, again, too early for this specifics, but I think OTI's commitment is to build out what that risk assessment process is and to be able to understand what we as the central technology agency for the city under E03, you know, what we need to do to be able to make sure that you know, tools are being developed and deployed responsibly.
1:06:32
Agencies as they do in other processes, other oversight processes, and review processes are a key player in that.
1:06:39
They're never it's never just OTI or or just the agency.
1:06:42
There's the agency is the business owner who's driving the need, driving the problem solving and the solution development, and then OTI providing support, whoever it can in in a a variety of different ways.
Jennifer Gutiérrez
1:06:52
So agency to agency in risk management or a risk assessment, excuse me, it will be up to that agency and OTI to figure that out.
Alex Foard
1:07:01
I it's a little too early to to know exactly what the process for this step by step, again, is because we're still in some of those earlier stages, like I described, better understanding that risk landscape.
1:07:11
It's
Jennifer Gutiérrez
1:07:13
Do you think that the results of of these assessments will be made public in the progress reports or just public by the the agency or OTI?
Alex Foard
1:07:21
Again, I think it's it's too early to know exactly what will be in that But transparency is one of your principles.
1:07:27
Transparency absolutely is, wherever we can.
1:07:29
Whoever we can, we like to make sure that information is available so that New Yorkers better understand the role that these technologies play.
1:07:36
I am not sort of under I I don't have the insight into exactly what we're asking in those questions right now to know exactly what would be available.
Jennifer Gutiérrez
1:07:47
But and so just for you in in your unit or department, Do you think that that is a direction?
1:07:56
The city, not just OTI, should move in where if where these these guidelines are to understand how to do risk assessment for your own agency.
1:08:07
It can be with OTI or without OTI.
1:08:09
It's too early to tell.
1:08:11
But that New Yorker should have access to this information.
Alex Foard
1:08:14
Yeah.
1:08:14
I think that's why we're so supportive of local law 35, which gives New Yorkers insight into the tools that are currently being used.
1:08:20
I think what is
Jennifer Gutiérrez
1:08:21
There's no I mean, I read it.
1:08:23
It's not yes.
1:08:25
It does do that, but it's it's not like I don't think it's explicit in, like, risk assessment.
1:08:31
It's not explicit yet on not all agencies respond equally, so there's, I think, a different level of detail agency tool to agency tool.
1:08:42
So that's why I think it I'm just asking if you think it should live separate in a in a part.
Alex Foard
1:08:47
Sure.
1:08:48
So just that
Jennifer Gutiérrez
1:08:49
The local output is an important role.
1:08:50
Yes.
Alex Foard
1:08:50
Yeah.
