AI Governance in Action: Insights from Austin's Local Government

0:00Kind of kicking it over, would you be able to tell everyone a little bit about what you're currently working on over at the City of Austin. And I guess if you could share some of the interesting ways that you are looking at applying AI to government work. So, yes. So, thank you, Jaden, for that introduction. So, I'm currently working as a senior data scientist at the city of Austin's Office of Innovation. We are a small team of different kinds of data sets and skill sets that we bring. So my role in the team is basically working on the quantitative side of research and analytics. So right now we are holding a lot of different projects where we are trying to improve the awareness of these newer technologies like data science, analytics, as well as artificial intelligence, and how it can be used, especially in context of local government.

0:54In my opinion, artificial intelligence and all of these emerging technologies are pretty crucial in making the government more transparent and accountable. So I always feel that we need to make the government services more efficient, more accessible to the public. And then it is also a great tool to improve how government communicate with the public. a very great example you are seeing a lot of chatbots at different places throughout the government settings as now as well so basically where you are seeing that the daily activities where people needs to go and maybe uh go go to a portal pay their bills now it is being used the chatbots are being used to automate that process making sure that it is effective and making it more accessible to the public so i very these technologies are pretty good very cool i do like that uh that's a that's a cool concept to be able to use like a chat bot to help you through your portal i know for a lot of things um government related like i'm just thinking of myself over here i believe it's through like the my government my city my city's government website that i like pay my um like electric and gas bills and that kind of stuff and sometimes there's really confusing things in there i can't figure out so having those chat bot uh experiences for that kind of stuff i think could definitely be like a beneficial to um customers, right?

2:18But also, I think this is beneficial to, you know, the city itself and being able to perhaps alleviate some of the work that customer service reps and that kind of have, those kind of people have. What are some of the unique challenges of implementing AI in a government setting, do you think? That's such a great question. So I think the two big things of implementing any AI product or technologies, in my opinion, the data should be reliable, and then data should be accurate. I always feel that these two things are the basic pillars of making any AI projects or any tools and technologies. So making sure that your data is reliable, you don't have any ethical issues connected with the data, the data should be up to date.

3:02And then it should be the format of the data should be in a particular way that any AI technologies can be used. so like data standards data governance and all of these things are pretty important especially from the data local government perspective because uh we have a lot of data so every city collects and owns a lot of data but what is important is to have that reliability and availability and making sure again going back to the first point that you raised that it should be transparent and the trust of the public should be maintained with whatever tools and technology we are building. So I think these are some of the great examples.

3:45So reliability, accessibility, I believe are like the two basic challenges of using any AI tool in a government setting. Okay. Yeah, that makes a lot of sense. Definitely some unique challenges there. You know, you mentioned talking about, you know, data management and whatnot. I'd be curious to pick your brain on like, how do you approach data management and, you know, pipeline development to ensure the success of AI projects at the city of Austin? So I think like the first step even on this data management should be to define the data strategy. So data strategy can differ based on the type of organization.

4:22So for city government or any kind of the government setting, I think it is important to have the principles and practices to ensure how data is connected, how it is stored, managed, and then used. So it has, you always have to consider those factors while you are doing any kind of data management. The second, for example, City of Austin has its own open data portal. They have a comprehensive inventory of different kinds of data assets City owns. So this is usually accessible by the general public. It has a lot, it has a big catalog which I can identify and manage all the data assets. you can go download, create some visualizations on your own.

5:08And then City of Austin especially owns a lot of different types of tools and technologies within the company which is being used to create better platforms like Tableau. We have Power BI. A lot of our housing department use ArcGIS. So all of these tools and technologies, I believe, are pretty important to make sure that we are having that data management aspect. And then the second point of your question on the pipeline development. Yeah. I think just making sure that we have the backend process intact. In maybe the latest stage of the podcast, I'm going to talk about a project which we recently did at Innovation Office where I can give you an example of how important it is to have that data pipeline stored.

5:54So again, a good team of data engineers, good mix of data engineers, data scientists, data analysts is very important to make sure that you have this lifecycle process intact in terms of any data project you have. Okay. Yeah, that makes a lot of sense. Specifically, something I'd be curious in asking you about is what do you kind of see the role that AI is playing in making governments more transparent and accountable, right? This is something you talked about a little bit at the beginning. You mentioned that it has this ability, but what role do you see AI really playing in that? So I think there can be a multi-faceted approach to this.

6:32So I gave you an example of a chatbot making the government services more accessible to the general public. That's an example. There can be a lot of natural language processing and algorithms that can be used to basically understand text and speeches. We get a lot of council member meetings. We get a lot of inside citywide meetings that are conducted so we can do a lot of those natural language processing there. Again, predictive analytics is a factor to understand and see what impact the different events can get. Another example, very local to Austin. We got a big no storm in 2021. No one was ready for it.

7:16So we got a lot of issues with that. how now how data can be used to understand the kind of impact and making sure that we are ready if there is any kind of such event in the future that can also be like a good example of how AI can be used a risk assessment again security making sure that we are ready for any kind of fraudulent or any kind of bad thing happening with the data as well as like in general the city services it cannot ai can be used there and then uh public access make it easier maybe i i in the recent uh the conference that i recently attended i saw a company where there was a device where it was like a glass which you can hear and then you can in real time see the translation of someone speaking to you in a different language by example if you are talking to me in English and let's say I am a native Spanish speaker and getting the real-time translation in front of my eyes which can help me in like better translating to you rather than me holding a phone in my hand asking you to say something, translate it, read it, and then speaking it.

8:28So I think these kinds of technologies are pretty great especially from the government perspective of how it can affect like the residents of the city. Very cool. Okay that's super cool. You know you You mentioned this just before this, talking about some specifics. I would love to hear them. I'm wondering if you could share maybe a specific project where AI has been, you know, has had a significant impact. And, you know, you mentioned like chatbots and whatnot, but perhaps there's other areas. I'm wondering if you have any interesting specific projects you've been working on. Yeah. So I can talk about a very recent project that we have done here at City of Austin.

9:05So I led a project called Neighborhood Prosperity Dashboard. which is basically a neighborhood level decision support tool that shows the current state of multiple prosperity related factors and again it is on a neighborhood level scale so pretty zoomed in understanding of how residents in different parts of Austin are thriving it incorporates an index of more than 80 different data variables across the five big categories economics, health, education, civic engagement, as well as social justice. So basically with that with that dashboard you get to see the state of neighbors and neighborhood so that different city departments can use it to operationalize their resources.

9:50It is being used a lot here at city of Austin to make a lot of important policy decisions. We are right now in the process of reaching out to different city departments and demonstrating it so that they can include it as as a part of their daily operations so they can do it use it to allocate resources or developing new services for the city and then the platform that we created going back to the initial point i said about the pipeline development the tool that we created uh holds a lot of data and then in the backend, it also is made with the fact that it can update whenever there is a new dataset available.

10:32Okay. And then in terms of reliability, we have used datasets from a lot of different data sources, census being one of one. We have some of our own local data. We got data from CDC's website. We got some data from the state of Texas, some data from the Travis County. So it's a collection of different types of data. and we are making sure that once the data becomes available, we are able to update the dashboard to show the latest information. Very, very cool. Wow, what an interesting project. Working on some of these really interesting projects, I'd be curious, what advice would you give to other cities looking to implement AI in their governmental operations?

11:12I think that's, again, that's a great question. I think the first part of starting with any kind of such projects is to have a clear understanding of your goals. So what are you trying to achieve with implementing, let's say, AI? Do you have clear understanding of your goals? And then do you have buy-in from your stakeholders? Stakeholders being both the city staff, the executive leadership, as well as the residents. You need that buy-in from all different parties. and then some suggestions that that I would make to the city start small you don't have to start with like a very big city-wide project start small create a prototype see if they have usability and then the second point is be transparent be transparent in front of your city residents or other like whatever whoever your customer is of what you are doing what kind of data you are using and what is the end goal of that project and then lastly but very important be responsible.

12:17I think you have to use AI in a responsible and ethical way. What does it mean that you have to be fair, unbiased, and transparent? And then you also have to make sure that you are protecting the privacy of the citizens. Yeah, that makes a lot of sense. And I think that is really, really good advice. Something I'd be curious on asking you about is, you know, as you're kind of working and implementing this, you're kind of on the forefront of AI in government work right now, I feel like you're doing some really impressive things. I'm wondering, what are some emerging trends in AI that you think are going to have a significant impact on government services in the next few years?

12:58I think hearing that question, the first thing that comes to my mind is generative AI. You are seeing it everywhere now. I think ChatGPT made a record of most number of users in a short amount of time ever. It is the most used app so again generative ai is the first one it is very transmute it is very transformative if you again use it ethically and use it in a responsible way it can improve your uh communication your document creation and service delivery what i usually tell to people who ask me question should i use a native way that uh you can use it but use it for the repetitive task and not for something where you have to back that with the data because, again, these tools and generative AI platforms are changing, but they are still lagging a lot in terms of, like, giving out the exact and current data set.

13:54So I would say for that, generative AI is great. Again, responsible AI is, again, an important point. So developing any AI model in a responsible and ethical way protect the privacy of your citizen, recognize the need of adopting these responsible AI practices. And then I'm thinking this is going to be a trend that is going to be here for some time in the next future as well.

14:24Another aspect or approach that I'm thinking is going to take over in the next few years is the multimodal AI. Basically, when you are going to bring images, text as well as like the structured data together and create some tools based on that so for example when you have like like an NPD for example that that neighborhood cross-party dashboard I gave you an example right now it is being backed by the data is what if we also bring the feedback from the community living actually living in those neighborhoods what if we can bring some kind of documentation some kind of text features feedback from people and try to implement it And chatbot is, again, an example of multimodal AI where you are bringing all the processes into a form of text-based software.

15:14So it can help you gain that understanding. And I would say this is going to be something which is being used a lot on across the domains, not only in government, but also in private organizations, too. That is so fascinating. It's been incredible to have you on the podcast. guys, you've shared some incredible insights, advice, and also some trends you see. If, you know, people are interested in contacting you, perhaps people that are also working on implementing AI into local governments or just other areas, what is a good way for people to, you know, reach out to you to find you? Oh, I think the best way is to reach me on LinkedIn.

15:53I'm here on LinkedIn at with my name. And then if you want to learn more about in general about what we are doing at City of Austin, I would suggest reaching out at innovation at austintexas.gov. So I-N-N-O-V-A-T-I-O-N at austintexas.gov. That would be a great place to start that conversation. And I would encourage all of your listeners to come and check what we are doing, give us the feedback, and then we can basically co-create any kind of product if you want. Amazing. So fascinating. Thank you so much for coming on the show today. And to the listeners, thank you so much for tuning into the AI chat podcast.

16:33Make sure to rate us wherever you listen to your podcasts and have an amazing rest of your day.