- The application of deep learning and generalized AI is still emerging in some U.S. public finance sectors such as local government, but has advanced in other sectors such as health care.
- There are several purposes of AI in the public finance space, including: reducing expenditures, alternative data pattern recognition, and public investment.
- State transportation agencies and health care entities are at the forefront of adopting AI in public finance.
- Many governments are using AI and machine learning to help address their cybersecurity needs.
The COVID-19 pandemic and the effects of climate change have accelerated U.S. government bodies' use of alternative data and the opportunities it provides to aid decision-making. Artificial intelligence (AI) tools are integral to its management and interpretation, although AI can mean many different things to different market participants. S&P Global Ratings credit analysts have had the benefit of talking to management teams around the country and in different sectors about technological adoption, which until recently, focused on creative ways to deliver services more cost-effectively through technology. The "Great Pause" has changed the conversation, and now management teams are collaborating with other governments and organizations to monitor and control the spread of the coronavirus, but also provide services virtually in lieu of the historically familiar in-person approach. Nevertheless, operational and budgetary challenges remain, and issuers are especially keen to identify innovative approaches to improve efficiencies and reduce costs in the face of recessionary pressure and operational model uncertainty. Many of these solutions have involved AI and digital improvements.
What Is AI?
AI is defined as any technique that enables computers to mimic human behavior. The term is often generalized to describe a wide variety of technologies, including machine learning, an AI subset which uses statistical models and deep neural networks to enable machines to improve through repetitive iterations, as well as deep learning. Each of these are attempts to imitate intelligent behavior, and usually involve the application of algorithms, which are statistical methods of interpreting data.
The application of deep learning and generalized AI is still emerging as organizations in some U.S. public finance sectors such as local government and health care work to understand AI's power and potential.
U.S. Public Finance Applications Of AI
One of the key questions related to AI in public finance is how entities will apply these technologies. With ever-growing demand for public services and budgets that are always stretched, prudent and innovative U.S. state and local governments identify opportunities to use AI to transform dated organizational models, rethink core operations, and reshape citizens' customer experience. Many local governments with long-term goals of reducing expenditures and improving citizen services have started incorporating AI into their day-to-day decision-making and strategic operations. We have identified several applications of AI in the public finance space thus far.
Health care triaging and outcome improvements
AI has the potential to disrupt the conception and delivery of health care, and never has that been more apparent than today. While AI was slowly being adapted to boost automation in health care, the COVID-19 pandemic could accelerate that trend.
Organizations are beginning to use AI technology to screen and triage patients, predict outbreaks, manage patient capacity, and devise treatment plans. While data access, reliability, and consistency remain challenges, the use of AI to track testing results and predict outbreaks has provided state and local governments with valuable data to decide what social restrictions to impose and subsequently lift. In fact, according to the magazine "Science," one of the first red flags regarding the pandemic came from HealthMap, a website run by Boston Children’s Hospital (AA/Stable) that uses AI to scan social media, news reports, and other information streams to identify initial signs of a disease outbreak. In late December, the program spotted a new report regarding a novel coronavirus in Wuhan, China and alerted other health care organizations about this report. This early alarm from HealthMap underlines the potential AI has to predict and therefore better prepare for future pandemics.
Hospitals are also realizing the value of AI technology when it comes to triaging patients and managing capacity. Providence St. Joseph Health System (AA-/Stable) in Seattle partnered with Microsoft to build an online screening and triage tool that was able to quickly differentiate COVID-19 and non-COVID-19 patients. The tool helped alleviate high volumes of patient traffic and allowed physicians to deliver care more effectively. The Massachusetts Institute of Technology (MIT) and its Computer Science and Artificial Intelligence Laboratory recently developed "The Emerald," a device that is able to remotely monitor patients and thereby prevent the spread of COVID-19 to health care workers. AI is also being used by several organizations to prioritize resources and research and thereby improve the search for COVID-19 treatments and vaccines. While AI's role in the pandemic is nuanced and still emerging, the application of AI technology is demonstrating its potential to transform health care organizations as it matures and adoption becomes more widespread.
Reducing expenditures: Transportation solutions
Municipal issuers and nonprofit entities, facing rising fixed costs, want to reduce expenditures through operational improvements and efficiencies, most notably through the adoption of innovative technologies. Inspired by private-sector success stories, government agencies have started adopting various AI technologies in diverse domains (e.g., health care, taxation, and education) and investing in technological infrastructure as well as software. For example, we have observed investments focusing on robust payment platforms to manage finances while facilitating greater transparency for disclosure purposes. Implementation of new technology usually features delays or unexpected transition issues, and we have seen budget problems and sometimes even the loss of data due to implementation errors.
A successful example of using AI involves Pittsburgh, Pa. (AA-/Stable), which collaborated with a private vendor to create automated traffic optimization and control software. City traffic control departments can use the software to manage traffic flows through several intersections and use AI to optimize the traffic light systems to reduce travel times, traffic stops, and wait times. The machine learning algorithm behind the software was fed labeled images of cars at various angles, lightings, speeds, and in various volumes of traffic. The system reportedly helped the city reduce travel time by 25%, traffic stops by 30%, wait time by 40%, and overall emissions by 21% during the course of the pilot. Such traffic management systems can aid in transit delivery reliability and generate growth in downtown areas, while reducing the costs associated with congestion.
States such as Maryland (AAA/Stable) and Vermont (AA+/Stable) have also invested in traffic management technology. Maryland has announced a $50 million upgrade of its aging traffic signals. The new signals can respond to traffic flow and conditions immediately, reducing travel times by 10%-15%. Vermont is modeling pavement conditions to predict (with 85% accuracy) how long a road treatment will last, predicting bridge deterioration using various models looking at over 100 different factors, and identifying sign locations to prepare for autonomous vehicles. These efforts to monitor traffic flow and infrastructure conditions will hopefully improve planning and driving time and cost efficiencies in the long term.
Improved communications: Bots are a form of AI
Municipalities can reduce expenditures by facilitating more effective citizen engagement. A major benefit of the new technology is that it enables the government to improve service delivery efficiencies, as well as help better manage the internal workforce. For example, chatbots use natural language processing to understand and respond to human requests. Los Angeles (AA/Stable) first introduced them in city government, followed by the State of Mississippi (AA/Stable), Kansas City, Mo. (AA/Stable), and San Francisco (AAA/Stable). Mississippi introduced a chatbot (aptly named MISSI) that can respond to over 100 enquiries, enabling the state to improve citizen engagement by addressing basic needs and enquiries much faster than before, thereby reducing or eliminating the usual congestion at service helpdesks. Through the use of chatbots, governments are trying to improve timeliness of service delivery and maintain the trust and support of their constituents.
Using Alternative Data For Climate Change Adaptation
As mentioned, obtaining new sources of alternative data is a major incentive for the application of AI. AI can be used to automatically identify patterns and understand local needs to thereby set and adjust policies. Of late, this has been most prevalent in understanding the physical effects of climate change, and other environmental, social, and governance (ESG)-related topics. For example, Louisiana (AA-/Stable), with over 7,700 miles of coastline, is constantly under threat from a variety of weather- and storm surge-related events. The region is home to large fisheries and petrochemical operations; in addition, Louisiana maintains five of the nation's top 12 ports by cargo volume and is home to nearly 2 million people. Many of the tax bases that support municipal obligations along the coast are concentrated in one of those industries. Damage from hurricanes and flooding has significantly altered the local ecosystem, with 1,800 square miles of land lost between 1932 and 2010, and 300 square miles alone lost from the recent Hurricanes Katrina, Rita, Gustav, and Ike.
As Louisiana communities and the state work to build coastal protection systems and improve drainage systems, researchers are using machine learning and neural networks (a type of machine learning that mimics the human brain and nervous system) to improve forecasts to help better understand the frequency and intensity of such environmental threats. Some coastal communities model storm surges, and with help from the Army Corps of Engineers, plan elaborate levee systems to withstand weather events. For example, the Terrebonne Parish Levee Conservation District, La. estimates that its levee system saved the region nearly $500 million in damages that the Hurricane Barry storm surge could have caused. Officials modeled storm surges, using data from previous storms to determine the appropriate dimensions and strength of the levee systems they needed. Other researchers, including groups at universities, NASA, NOAA, and USGS, continue to use machine learning and neural networks to improve forecasts, hoping to generate more accurate predictive models.
Other examples include the Minnesota Pollution Control Agency, which used AI and analytics to improve real-time weather information to allow vulnerable populations to receive weather alerts in the event of an emergency to facilitate preparedness actions. Also, AI researchers from Georgia Institute of Technology, Emory University, and the University of California at Irvine worked with the Atlanta Fire Rescue Department (AFRD) to develop a predictive analytics software aimed at identifying buildings that have a higher likelihood of fire incidents. The researchers developed Firebird, which purportedly uses historical data for 58 variables such as property location and building descriptors made available to them by AFRD. The data were fed to an AI predictive analytics software and the algorithms were tweaked to forecast fire risk scores for 5,000 buildings. According to the university, the software accurately predicted 73% of fire incidents.
From a broader credit perspective, geospatial data are essential in identifying specific trends. For example, by linking nature to debt metrics, S&P Global Ratings identified that utilities in regions with evergreen forests and perennial ice and snow had greater all-in coverage ratios than those located elsewhere (See "Space, The Next Frontier: Spatial Finance And Environmental Sustainability," published Jan. 22, 2020 on RatingsDirect).
From a credit perspective, by applying these technologies, we believe local governments can stay ahead of potential risks and prepare accordingly.
Investment In The Public Arena
Finally, while not technically a direct application, we have noticed issuers across the country investing in AI and technology, whether for workforce development purposes or for cybersecurity.
The ability of government organizations to attract and retain cybersecurity professionals somewhat reflects pay-grade structures and competition from the private sector. As the amount of stored data grows and the available qualified workforce fails to keep up with demand, AI security software, which identifies abnormalities in network and software functionality, can detect threat patterns and allow the professionals to deal with most necessary human-led tasks. Software that includes machine learning may be able to automatically respond to the growing sophistication of attacks and learn new patterns. This is further strengthened by public finance entities improving their own cybersecurity awareness and buying insurance coverage, either independently or through larger regional plans. While action in this area is urgent, disclosure of relevant actions can put issuers at risk. (See "U.S. Public Finance Issuers Must Be Nimble To Fend Off Cyberattacks Or They Could Face Credit Fallout," published Feb. 25, 2020.)
New Orleans (AA-/Negative) was targeted in a December 2019 cyberattack, which significantly impaired the city's operations in areas such as revenue collection, law enforcement, and judicial courts. Currently, the city expects costs from the intrusion to exceed $7 million, but officials credit investments in cloud-based solutions for financial systems for mitigating the damage. New Orleans plans on further investments to harden IT systems and increasing cyberinsurance coverage.
Beyond the risks of cybercrime, several issuers are developing the workforce to better exploit these technologies. Some states, such as North Carolina (AAA/Stable), have a dedicated mandate to improve the workforce across all sectors, but particularly with regard to technological and IT skills. As a result, it has encouraged its municipalities to partner with local colleges and universities to ensure training is available to the broader public in line with local economic development goals. Furthermore, since many of these ideas and innovations can be scaled to have a more global effect, major universities such as MIT are partnering not only with local cities (e.g., Boston and Cambridge), but also with a consortium of multidisciplinary partners to explore various AI solutions. Seeking local solutions to global problems enhances the benefits of AI technologies.
Challenges And Opportunities
The spectrum of wealth across public finance entities creates analogous challenges and opportunities in AI. While some larger and richer entities can make room in their budgets for the latest technological advancements and research, others are fiscally challenged, which hinders the wider and more rapid adoption of AI across the sector as a whole. Furthermore, appropriate governance and reliable data are needed to ensure that AI capabilities are fully understood and used responsibly. However, the diverse American landscape with differences among states means that data might be either lacking or incomparable. The challenge is global, as data reliability is an important issue nearly everywhere.
At the same time, we have noticed that some public entities are attempting to overcome these obstacles. Notably, due to the COVID-19 pandemic, educational facilities and school districts have now been required to expedite the provision of online learning, in itself an application of AI for some features. They are now the best-prepared institutions across the U.S. in this arena since the transition to blended learning models was already underway. Furthermore, a nationwide emphasis on science, technology, engineering, and mathematics (STEM) education has been a catalyst for several state and national initiatives that focus on staff and workforce development to meet the labor demand of the growing AI sector. At the same time, government entities are finding credible partners to share best practices, hiring third parties for their IT management, and collaborating in innovative research and urban design across multiple sectors. Given these efforts in a world that is rapidly changing, the full credit effects of AI and its challenges are yet to be determined. However, we believe those who make an effort to embrace the technologies will move the needle, whereas those who do not will be left behind.
- Through The ESG Lens 2.0: A Deeper Dive Into U.S. Public Finance Credit Factors, April 28, 2020
This report does not constitute a rating action.
|Primary Credit Analyst:||Kaiti Vartholomaios, New York + 1(212) 438 0866;|
|Secondary Contacts:||Alex Louie, Centennial 303-721-4559;|
|Aamna Shah, San Francisco (1) 415-371-5034;|
|Geoffrey E Buswick, Boston (1) 617-530-8311;|
No content (including ratings, credit-related analyses and data, valuations, model, software or other application or output therefrom) or any part thereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of Standard & Poor’s Financial Services LLC or its affiliates (collectively, S&P). The Content shall not be used for any unlawful or unauthorized purposes. S&P and any third-party providers, as well as their directors, officers, shareholders, employees or agents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Parties are not responsible for any errors or omissions (negligent or otherwise), regardless of the cause, for the results obtained from the use of the Content, or for the security or maintenance of any data input by the user. The Content is provided on an “as is” basis. S&P PARTIES DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT’S FUNCTIONING WILL BE UNINTERRUPTED OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no event shall S&P Parties be liable to any party for any direct, indirect, incidental, exemplary, compensatory, punitive, special or consequential damages, costs, expenses, legal fees, or losses (including, without limitation, lost income or lost profits and opportunity costs or losses caused by negligence) in connection with any use of the Content even if advised of the possibility of such damages.
Credit-related and other analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed and not statements of fact. S&P’s opinions, analyses and rating acknowledgment decisions (described below) are not recommendations to purchase, hold, or sell any securities or to make any investment decisions, and do not address the suitability of any security. S&P assumes no obligation to update the Content following publication in any form or format. The Content should not be relied on and is not a substitute for the skill, judgment and experience of the user, its management, employees, advisors and/or clients when making investment and other business decisions. S&P does not act as a fiduciary or an investment advisor except where registered as such. While S&P has obtained information from sources it believes to be reliable, S&P does not perform an audit and undertakes no duty of due diligence or independent verification of any information it receives. Rating-related publications may be published for a variety of reasons that are not necessarily dependent on action by rating committees, including, but not limited to, the publication of a periodic update on a credit rating and related analyses.
To the extent that regulatory authorities allow a rating agency to acknowledge in one jurisdiction a rating issued in another jurisdiction for certain regulatory purposes, S&P reserves the right to assign, withdraw or suspend such acknowledgment at any time and in its sole discretion. S&P Parties disclaim any duty whatsoever arising out of the assignment, withdrawal or suspension of an acknowledgment as well as any liability for any damage alleged to have been suffered on account thereof.
S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respective activities. As a result, certain business units of S&P may have information that is not available to other S&P business units. S&P has established policies and procedures to maintain the confidentiality of certain non-public information received in connection with each analytical process.
S&P may receive compensation for its ratings and certain analyses, normally from issuers or underwriters of securities or from obligors. S&P reserves the right to disseminate its opinions and analyses. S&P's public ratings and analyses are made available on its Web sites, www.standardandpoors.com (free of charge), and www.ratingsdirect.com and www.globalcreditportal.com (subscription), and may be distributed through other means, including via S&P publications and third-party redistributors. Additional information about our ratings fees is available at www.standardandpoors.com/usratingsfees.
Any Passwords/user IDs issued by S&P to users are single user-dedicated and may ONLY be used by the individual to whom they have been assigned. No sharing of passwords/user IDs and no simultaneous access via the same password/user ID is permitted. To reprint, translate, or use the data or information other than as provided herein, contact S&P Global Ratings, Client Services, 55 Water Street, New York, NY 10041; (1) 212-438-7280 or by e-mail to: email@example.com.