QUESTION
What is the role of the Mayor's Office of Data Analytics (MODA) in the New York City Artificial Intelligence Action Plan?
0:20:24
·
139 sec
Martha Norrick, Chief Analytics Officer, explains that the Mayor's Office of Data Analytics (MODA) plays a critical role by ensuring the use of high-quality data and comprehending how data are generated.
- Advances in AI and machine learning depend on good data and understanding the generation of data.
- The example of 311 calls about rats illustrates the limitations and biases in data sets.
- Understanding the nuances of data generation is essential for effectively leveraging AI technologies.
- Data scientists at OTI are instrumental in ensuring appropriate data usage in AI-government tool advancements.
Jennifer Gutiérrez
0:20:24
How do you see Moda's role in New York City Artificial Intelligence Action Plan announced by Mayor Adams?
Martha Norrick
0:20:32
The way I think about our role is that all of the advances in artificial intelligence and large language models in in these these sort of, you know, both new and not new applications of of machine learning and artificial in intelligence really depend on good data and also understanding how data are generated.
0:20:53
So the example that I like to talk is 311.
0:20:56
Right?
0:20:56
Like, how many people in this room have seen a rat and not called 311?
0:21:00
Great.
0:21:01
Okay.
0:21:01
Every every day.
0:21:03
Sometimes, you know, but we we people sometimes use 311 calls about rats as sort of a proxy for for, you know, understanding where rats are in the city, but it's really it's a it's a that data are generated by a by a process of of of New Yorkers calling in about something.
0:21:21
And we know that New Yorkers do not call in about everything they see, and we know that they don't call in about things that they see, you know, at equal rates, a across the city.
0:21:31
So if you're using the 311 data set, which is a huge, you know, people think it's so big, so it must be, you know, it must be perfect for everything, you know, the bigger the data, the better.
0:21:42
And and don't really understand sort of how that data is generated and how how to understand, you know, some of the biases or some of the gaps that that dataset contains, not because the data is bad, but because the way that it's made is that people call through 11, and not everybody does that and not everybody does that at every you know, for every for every rat they see in the city.
0:22:03
So I think really kind of, you know, we sit very close to, you know, we use this data.
0:22:09
We Our data scientists our data scientists are sort of navigating these same questions every day about what is the data set, you know, what type of questions is the data set appropriate to answer.
0:22:19
If we're using this dataset to answer this question, what would be what would be be missing?
0:22:24
And I think sort of that knowledge and that under standing of kind of, you know, end documentation around the data and what it means and what it doesn't mean and what you can use it for and what you should be careful about.
0:22:35
Are really going to be fundamental to to the AI, the advancement of of of using these new tools throughout government.
