Transformative Era: Artificial Intelligence within Local Governments
In the technology information era, according to McKinsey, local governments will be able to effectively use emerging technologies at various scales (Hieronimus). Local governments should harness AI innovations as a tool to meet their mission of communicating efficiently and effectively with residents. Let’s consider the City of Glendale; we were one of 12 cities selected across North America to receive free innovation training from Bloomberg Philanthropies due to our continuous efforts to remain up to date (Innovation). With a human-centric focus, putting residents at the center of their work, Glendale is known for its innovation and ways to allow easier accessibility for residents. Yet, have we taken advantage of AI? In this article, we will explore ways local governments, like ours, can utilize AI tools to manage internal and external communications and operations.
Given that local governments should maintain open and effective communications, Glendale should leverage emerging technologies. Artificial Intelligence will enable the City of Glendale to increase the speed and clarity of its communication with the public, but leadership must ensure that clear regulations are standardized to identify and mitigate risks.
Internal Communications:
Various local governments began adopting ChatGPT to improve internal operational communication efficiency. Within the City of Glendale, the Communication Teams prepare memos and press releases that present analysis, findings, and recommendations concerning policies, research, and adoption of ordinances. When given the correct prompt, ChatGPT can generate templates for memos and press releases that can later be tailored by city staff with the correct information about upcoming publications. ChatGPT will prevent continuous repetitive work and according to McKinsey, repetitive manual work is frequently cited as one of “the main sources of public-sector job dissatisfaction” (Hieronimus). To further improve communication strategies, the start is often the most time-consuming part of writing. In the prompt, city staff can include the basic details of the problem and how city officials are responding, including a generic “quote” from executives (Rueter). Afterward, the draft can be edited by one of our city staff, and replace the quotes with correct information, enhancing communication speed externally and internally.
Exhibit 1: Improving internal operations in local governments.
Source (McKinsey)
From a city executive perspective, they read lengthy reports and make important decisions quickly. To speed up the process, for example, Singapore developed an app that summarizes text and generates reports for internal use (Bruce). The same application is relevant to the City of Glendale, where we can summarize lengthy memos and reports with all of the essential claims and analyses (see exhibit 1). Suppose a city official wants to make a decision based on a summary. In that case, they must still read the entire document to ensure they did not miss or mischaracterize the original document (Garces). In return, city officials will be able to learn uprising information and support any internal conflicts quickly.
External Communications:
In addition to supporting our internal communications, generative AI can also enhance our external communications, including during critical periods such as when disasters strike and residents panic. During disastrous times, residents begin to look to local governments for support and active communication. Recently, we experienced Hurricane Hillary, where panic rose quickly within residents. Within 24 hours, the executives expected our communications team, consisting of 3 individuals, to provide timely media updates and press releases. Had our office had an LLM protocol in place at the time of this disaster, we would have been equipped to meet our residents’ demands by creating written drafts that instantly respond to the panicking residents
The AI-enabled emergency notification system can take weather data from the National Oceanic and Atmospheric Administration, and create a series of alert messages in advance to warn residents about upcoming emergencies or crises (Haseley). The communications team can use AI to provide information in advance to areas that might have been affected more severely. Knowing that residents are looking for various disaster assistance programs, AI can “act as a single front door to various programs” and ensure every resident has received assistance in any form (Haseley). Besides increasing the speed at which information is distributed via ChatGPT, how can AI assist with the clarity of public communication through local governments?
ChatGPT can facilitate communication in multiple languages, ensuring inclusivity and accessibility for all residents. According to the Census, Glendale consists of 13.7% Asians, 18.9% Hispanics, and 68.6% white, which is mainly Armenians (U.S. Census). Within this population, around 66.7% speak a language other than English in their household, which translates outreach and communications is important within the City of Glendale. Translating in appropriate languages would help the residents have easier access to various services. Sites like Worldly provide AI-generated translations that can assist during City Council Meetings, Town Halls, and more! Upon translation, the format, tone, and key messaging are all factors that need to be considered when preparing and approving written documentation (Garces). Although this increases accessibility, AI-generated translations may not always be accurate, which is why there must be on-staff individuals who speak those languages to ensure accuracy. Similarly, AI-generated translations might be confusing to residents who speak different regional dialects (Garces). This is where understanding how to develop the prompts that target specific regional dialects becomes crucial with the support of a department staff.
Exhibit 2: Customer Facing Operations in Public Sectors
Source (McKinsey)
The City of Glendale’s recent project of introducing digital signatures using an automated process shows their commitment to innovation to speed communication with residents (Innovation). The next step is to use ChatGPT to enhance the maintenance of infrastructure and public services. In neighboring cities, for example, Los Angeles’s Department of Transportation is using “AI-powered chatbots to provide real-time traffic updates’’, which in return helps them plan their routes and ways to avoid traffic congestion (Bruce). This is just one of the many examples; others include waste management, urban planning and public safety cases as showcased in Exhibit 2.
Taking this into account, the City of Glendale can take advantage of AI-powered chatbots to answer common questions and provide generic services. Local governments receive daily calls and emails asking about permits, bulky item pickups, power outages, and more. With supporting residents being the main goal, hours are spent trying to assist everyone with their basic questions. One solution would be to use Chatbots, which will provide real-time access for residents to receive in-depth answers that are tailored to their needs (Bruce). An expected city staff response to a resident’s inquiry is to direct them toward various links (Freed). Chatbots will now provide summarized responses with appropriate information. Internally, it will reduce the workload on human responders and free them up to focus on more complex tasks (Freed). However, residents may start questioning how accurate these AI responses may be and their setbacks.
In order to prevent inaccuracy, the City of Glendale needs investments in AI research and innovation and set clear regulations. According to McKinsey, lessons from the private sector show automation at scale has the potential to serve those purposes. Local governments must become more strategic in their approach and be prepared to act at a scale (Hieronimus). One notable piece of information is to educate the residents and the staff that ChatGPT cannot determine accuracy or understand bias. When using a free AI account it’s important to know that all of the inputs provided belong to the company that made the AI-generated tool. They have the right to use that information, thus it is available to the public.
Recently, the City of Boston generated an interim AI guideline for their staff. Their initial step was to educate the team that “[they] are responsible for the outcomes of [the] tools” used (Garces). For any changes and alterations to words and meanings, city staff must proofread and adjust those changes accordingly before anything is presented to the public. In these guidelines, they listed the do’s and don’ts of how a city should use ChatGPT. For example, they don’t allow staff members to “input private data or personal data, such as emails, birthdays, or social security numbers, into ChatGPT or similar AI-powered tools” (Garces). This addresses many concerns that the public might have about data and security protection from the government’s end. When using any AI-generated tools, local governments can prevent data and security breach issues by closely monitoring their data inputs and using a paid account.
A lot of the information that is public facing is already public, which is why it can be easier to use ChatGPT for press releases and media management. The City of Boston has taken this into consideration and provided an alternative that can showcase transparency between local government and its residents by disclosing that they have used AI to generate the content. At the end of the writing, city staff can include “This description was generated by ChatGPT 3.5 and edited by [Mariam Hagopian]” (Garces). This may be obvious to many but some may be against disclosing that information to maintain their credibility and reputation of being a resourceful tool for residents. Instead, this provides the residents with the understanding that some inaccuracy might occur and in that case, they can contact their local city for further information. This type of disclosure can be used both internally and externally, helping others catch errors (Garces). We can easily implement City of Boston’s interim regulations with minimal financial spending.
With our efforts to adopt ChatGPT to increase the speed and clarity of its communication with the public, they will have to provide mandatory staff training about the regulations, risks and ways to write prompts that can later be developed as a city regulation for everyone to use. In this article, we explored various ways local governments, specifically City of Glendale can utilize ChatGPT through operational, emergency and public service perspectives. Generative AI is not the solution to all the problems that exist in local governments, but it is a took for effective communications both internally and externally. There is more to consider over the next couple years as Large Learning Models are evolving and changing the ways regular businesses function.
Works Cited
Bruce, Damien, et al. “Unlocking the Potential of Generative AI: Three Key Questions for Government Agencies.” McKinsey & Company, 7 Dec. 2023, www.mckinsey.com/industries/public-sector/our-insights/unlocking-the-potential-of-generative-ai-three-key-questions-for-government-agencies. Accessed 29 Jan. 2024.
Chui, Michael, et al. “The State of AI in 2023: Generative AI’s Breakout Year.” McKinsey & Company, 1 Aug. 2023, www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-ais-breakout-year. Accessed 29 Jan. 2024.
Dukowitz, Zacchary. “AI for Government: Chatgpt in the Public Sector [with Prompts!].” OpenGov, 12 Sept. 2023, opengov.com/article/ai-for-government/. Accessed 29 Jan. 2024.
Freed, Benjamin. “CHATGPT Said It Will ‘Revolutionize’ Digital Government. Human Experts Are More Cautious.” StateScoop, 5 Apr. 2023, statescoop.com/chatgpt-government-digital-services/. Accessed 29 Jan. 2024.
Garces, Santiago. “_Guidelines for Using Generative Ai.Pdf.” Google Drive, Google, drive.google.com/file/d/1o25MruYz8NYgJNLglfcrGUhr8m7BJAkE/view. Accessed 29 Jan. 2024.
Haseley, Alex, et al. “Leveraging AI for Effective Emergency Management and Crisis Response.” Deloitte Insights, Deloitte, 1 Dec. 2023, www2.deloitte.com/us/en/insights/industry/public-sector/automation-and-generative-ai-in-government/leveraging-ai-in-emergency-management-and-crisis-response.html/#:~:text=For%20instance,%20an%20AI-enabled,an%20impending%20hurricane%20or%20other. Accessed 29 Jan. 2024.
Hieronimus, Solveigh, and Navjot Singh. “McKinsey on Government Perspectives.” McKinsey on Government Perspectives: Adopting AI, Automation, and Advanced Analytics in Governments, 5 May 2019, www.mckinsey.com/~/media/McKinsey/Industries/Public%20Sector/McKinsey%20on%20Government%20May%202019%202/McK-on-Government_5-v2.pdf. Accessed 29 Jan. 2024.
Innovation | City of Glendale, CA, www.glendaleca.gov/government/departments/management-services/office-of-innovation-performance/innovation. Accessed 29 Jan. 2024.
Mascho, Brad. “Unleashing the Power of Chatgpt in Government.” Federal News Network — Helping Feds Meet Their Mission., 30 May 2023, federalnewsnetwork.com/commentary/2023/05/unleashing-the-power-of-chatgpt-in-government. Accessed 29 Jan. 2024.
Rueter, Thad. “Can CHATGPT Write Better Press Releases for Municipalities?” GovTech, 3 Mar. 2023, www.govtech.com/biz/can-chatgpt-write-better-press-releases-for-municipalities. Accessed 29 Jan. 2024.
U.S. Census Bureau Quickfacts: Glendale City, California, www.census.gov/quickfacts/fact/table/glendalecitycalifornia/INC110222. Accessed 29 Jan. 2024.
