S&P Global Offerings
Featured Topics
Featured Products
Events
S&P Global Offerings
Featured Topics
Featured Products
Events
S&P Global Offerings
Featured Topics
Featured Products
Events
S&P Global Offerings
Featured Topics
Featured Products
Events
Language
Featured Products
Ratings & Benchmarks
By Topic
Market Insights
About S&P Global
Corporate Responsibility
Diversity, Equity, & Inclusion
Featured Products
Ratings & Benchmarks
By Topic
Market Insights
About S&P Global
Corporate Responsibility
Diversity, Equity, & Inclusion
Smart cities could become even smarter through increased application of AI. Learn what AI smart cities are and how they can impact people's lives.
This is a thought leadership report issued by S&P Global. This report does not constitute a rating action, neither was it discussed by a rating committee.
Published: 18 May 2024
By Zoe Roth and Melissa Incera
Highlights
Smart cities could become even smarter with increased application of AI, both in infrastructure development and analysis of data.
Enhancing safety, sustainability, quality of life and resident experience are key benefits of urban areas becoming more AI-powered.
However, risks such as data privacy and critical infrastructure security are on the rise, and add to the challenge for governments and public bodies to govern and contain.
Over the next five years, we expect AI/generative AI to impact cities through integration into digital government services, smart transportation and interactive digital twins.
Cities are increasingly at the confluence of the world’s most pressing issues: dealing with extreme weather events, managing migration, maintaining affordability and ensuring public safety. As the share of the population living in cities continues to grow, leaders are looking to technology to help them solve some of their most pressing issues.
While many narrow their definition of smart cities to ones characterized by flying cars and advanced robotics, the reality is that any city or town that has added sensors to its infrastructure, including something as simple as streetlights, is already a smart city. By instrumenting critical systems with connectivity and adding sensors to existing infrastructure, cities can better understand their resource utilization, citizen behavior and service gaps. The advent of AI, and more recently, generative AI, has opened up a world of possibilities for expanding access to services and enhancing efficiencies. However, these advancements also create new risks.
As we navigate the transition to smarter cities, it is imperative for policymakers, industry stakeholders and communities to work collaboratively in addressing these challenges and ensuring that technological innovations are deployed ethically and inclusively. But by embracing a forward-thinking approach that prioritizes sustainability, equity and resilience, AI smart cities could usher in a new era of urban prosperity and well-being for generations to come.
Smart cities are cities, towns and communities that rely on the implementation of information and communications technology (ICT) and infrastructure, including artificial intelligence/machine learning (AI/ML) and internet of things (IoT) technologies, to derive actionable insight from their existing infrastructure, systems and processes to enhance quality of life and safety for citizens. City drivers for adopting IoT and other smart city initiatives vary depending on geography and size. Whereas some cities are driven by enhancing public safety, others focus on sustainability improvements, quality-of-life enhancements, or other operational efficiencies around resource utilization. Smart city applications appeal to citizens across rural and urban cities and towns (see Figure 2).
The smart cities movement has waxed and waned in the last decade in line with concerns around privacy and security, and more recently, pandemic re-prioritizations. Whereas the movement was initiated by large tech companies looking to offer platform-as-a-service or city-as-a-service approaches (e.g., GE’s CityIQ, Google Sidewalk Labs, Microsoft CityNext), cities have since pulled back their procurement processes to focus on outcomes and establish data governance structures to ensure longevity of projects.
Many cities are focused on transportation-related use cases, including public transportation optimization, intelligent intersections and multimodal transit network improvements. Some cities have taken a step back from larger-scale IoT deployments in favor of building out their connectivity infrastructure, including fiber networks, to ensure the longevity of initiatives. Broadly, a city will often focus on topics like sustainability and safety enhancements, and then work backward to determine technologies that may help them achieve those outcomes. Many deployments begin with small-scale, block-level pilot projects before reaching scale.
AI comes into play in smart cities when AI capabilities are applied to existing or new datasets or streams. Whereas IoT applications collect the data, AI analytics can detect patterns, make predictions, unify data streams (data fusion) and enhance data quality. When examining the AI + IoT (AIoT) equation, use cases that emerge vary between applications and cities. Many IoT application providers offer AI analytics as part of the insights they deliver to customers, offering historical analysis and prescriptive insight into metrics like planned resource utilization and demand forecasting. AI governance, data quality and employee data literacy are chief considerations to make before procuring and deploying AI applications and a full-scale AI strategy within a city.
AI smart cities are underpinned by pervasive digital infrastructure including connectivity and compute resources to support the collection and analysis of data. Connectivity can vary, though most cities rely on some combination of cellular networks (4G/LTE and emerging 5G), Wi-Fi (municipal or private-owned) and wide area networks (LPWAN/LoRaWAN). These networks provide the data highway between information-collecting endpoints and the cloud, where data is often processed and acted upon.
Where data is processed depends on the use case. Mission-critical or transportation-related applications, such as intelligent traffic signals, rely more heavily on edge computing (where decisions are made close to the data source) and rapid processing of data. A smart streetlight, however, has little need for high-speed data. Whereas both are enabled by sensors, an intelligent traffic signal may collect and process a signal from an onboard unit (OBU) in an emergency response vehicle a mile away and relay its location to a roadside unit in a traffic cabinet.
After receiving that signal, the edge unit may run algorithms to determine how to best direct the emergency vehicle through the intersection while minimizing downstream effects. A smart streetlight, on the other hand, may send one small data packet per day on energy used. The ultimate goal of an AI smart city is to derive actionable insights from processes and systems to enhance efficiency and create a demand-responsive city.
Cities globally have begun to establish smart city offices and directorates with dedicated staff resources and budgets to support projects and initiatives. Countries have established smart city challenges to spur innovation, and allocated funding to support project establishment and scale. These include the USDOT Smart City Challenge, the EU’s Intelligent City Challenge and India’s Smart City Mission.
Who “owns” smart city initiatives varies from city to city and country to country. Whereas some cities may have a smart city director or manager, others may work with a corporate partner to spur projects. Progress of the smart city movement can be measured in the successes of individual cities, such as Barcelona or Hong Kong, or in the increase of connected devices as a whole, extending globally to cities and towns of all sizes.
Cities adopt IoT and emerging technologies to enhance safety, sustainability, quality of life and resident experience. According to 451 Research's Voice of the Enterprise: Internet of Things, the OT Perspective, Use Cases and Outcomes 2023, 50% of government respondents selected ensuring public safety as the main driver for their smart city initiatives, followed by improving overall quality of life (44%) and improving city services (42%).
Safety (video surveillance, smart streetlighting, gunshot detection): Cities turn to vision- and audio-based technologies to enhance safety and improve incident response times. While surveillance-based or “safe city” initiatives can raise privacy concerns, cities can address these by working with vendors that follow "privacy by design" principles, as well as by establishing data retention policies and engaging the community early and often regarding camera-based deployments. Las Vegas has deployed smart parks that collect situational awareness data, including activity when parks are closed. Optical sensors and movement analytics create automatic alerts for public safety teams.
Sustainability (digital twin, waste management, public transportation improvement, water quality monitoring): Sustainability is front of mind for city leaders as they manage resource utilization, migration and climate events. Cities such as Lisbon have leveraged digital twins to better prepare for urban flooding by modeling what areas of the city may be most susceptible. With a vulnerability map, the city can better prepare and mitigate flood impacts when weather events occur.
Quality of life (free public Wi-Fi, public transportation improvements): Technology can enhance day-to-day experience for citizens by improving their commutes, air quality, and the services they receive from their governments. London, widely recognized as one of the smartest cities in the world, takes a people-centric approach to smart city deployments and has focused on facilitating pervasive digital access (including fiber and 5G connectivity). Atop this digital infrastructure, the city leverages connected assets and sensors to provide services like hyperlocal air quality data and contextual maps that can layer multiple datasets, like financial inclusion data by neighborhood.
Resident experience (open data portals, digital government services): Improving citizens' interactions with government can enhance satisfaction and optimize employee resources. By offering digital government service portals, such as an online permitting platform, governments can reduce barriers for their citizens. Philadelphia has launched a permit navigator pilot program that provides information and estimates costs for residential and business permit projects.
In smart transportation and intelligent transportation systems (ITS) examples of AI data analytics are pervasive across applications. Whether it is intelligent routing software providing optimized routes for an on-demand public transit service or predictive analytics for road safety, machine learning algorithms play an instrumental role as the “brain” of smart city applications. The benefits of AI analytics applied to transportation are multiple, from reducing traffic congestion to providing predictive analytics about accident-prone roadways or intersections.
Beyond pure analytics, computer vision (CV) enables many smart city outcomes related to public safety and efficiency. In computer vision deployments, images and videos are analyzed using algorithms (often at the edge) to provide image understanding/recognition, prediction, segmentation and reconstruction. Computer vision can enable smart parking zones and automated license plate recognition (ALPR) among other use cases.
CV can free up government resources by deriving insights from lengthy video streams or real-time video and relaying metadata or insights back in visual reports or dashboards. The long-term reliability of computer vision models depends on specialized infrastructure and expertise for model-building strategy.
Generally, cities are warming to the idea of artificial intelligence, though they recognize there are risks. For example, some police departments are using CV combined with other citizen data, such as previous criminal records, for predictive policing, which means that AI predicts when and where crime is likely to occur. This type of AI use case raises concerns of embedded gender, ethnicity and cultural bias hidden in AI models.
Every technological advancement comes with concerns about threats and unintended consequences that may arise from its adoption. Within the context of cities and government agencies, these risks are magnified due to the critical role public services play and the sensitive nature of the data involved.
For instance, in a recent 451 Research survey, government sector respondents expressed heightened concerns regarding the reliability and security of AI infrastructure compared to respondents across industries (see Figure 3), underscoring the need to build robust safeguards and protocols to uphold the integrity of smart city initiatives. These concerns include:
Data privacy: The extensive deployment of sensors and video-based surveillance systems in AI smart cities leads to the collection of vast amounts of personal data. This data encompasses individuals' movements, behaviors, preferences, and interactions with urban infrastructure. The aggregation of such sensitive information raises significant privacy risks, as citizens may feel uneasy about the constant monitoring of their activities in public spaces and the potential for data misuse or exploitation by governmental or corporate entities. Several notable smart city projects in the past have shuttered due to community pushback on surveillance, data privacy and transparency on how the collected data would be used.
Security and cyber resiliency: The interconnected nature of smart city infrastructure, comprising networks, sensors and communication systems, creates opportunities for malicious actors to exploit vulnerabilities and launch cyberattacks. These attacks could range from data breaches and identity theft to ransomware attacks targeting critical infrastructure, such as transportation systems or utility grids. The consequences of such cyberattacks may be far-reaching, posing risks to public safety, economic stability and the uninterrupted delivery of essential services. The public sector has significant surface area for cyberattacks, potentially correlating to trillions of dollars in direct and indirect loss. While direct costs may capture ransoms or regulatory fines, indirect losses including downtime of critical infrastructure, legal liabilities and reputational damage may be much more significant.
Digital equity and bias: Additionally, the emergence of a tech divide poses a threat to equality within AI smart cities, as older or lower-income segments of the population may struggle to adapt to smart city initiatives or be underserved by high-speed connectivity . These disparities in access to technology and digital literacy may widen the gap between those who have access to advanced digital services and those who do not, highlighting the importance of inclusive design and equitable deployment strategies to ensure that smart city initiatives benefit all citizens, regardless of age, income, or technological proficiency.
Addressing these risks will require proactive measures, including proper system design and guardrails for safe data usage. Despite the high enthusiasm for smart city initiatives, cautious adoption is warranted, with governments opting for a slow and deliberate approach to avoid potential pitfalls.
The near-term impact of AI on smart cities will likely center on the application of traditional AI, with generative AI mixed in. We expect that state- or city-level guidelines on risk mitigation and procurement, outcome-driven use-case analysis, vendor vetting, and sharing of best practices between public sector organizations will precede AI adoption. A lot of this work in the U.S. is being driven by the San Jose-led Government AI Coalitions, which unifies public sector leaders to plan for, procure and implement AI and generative AI. Key areas where we expect to see AI’s impact in smart cities are enhanced digital government services, smart transportation integration, and digital twin/city modeling.
Digital government: AI will support government’s ability to deliver enhanced, personalized digital government services to residents. These services can include digital IDs, online services such as permitting and filing taxes, and support for electronic healthcare services. Tools that can enhance government employees productivity in generating and analyzing documents and other content will play a key role in enhanced digital government.
Estonia has established a mature model for a digital society, enabled by government services that support nearly all aspects of daily life from schooling to banking and telehealth. Under the country’s AI strategy for digital services, AI is supporting Bürokratt government services chatbots, AI-driven tax fraud detection, computer vision in resource monitoring, automatic translation and subtitling for government events, and other use cases.
Governments will turn to AI, namely service chatbots, to free up employee resources and better serve constituents, with an emphasis on adopting trustworthy, explainable AI applications. Nonetheless, these new initiatives can prove problematic. New York City’s Azure AI-powered chatbot gained headlines recently when it repeatedly shared inaccurate information, such as telling landlords they could reject tenants with Section 8 vouchers, and informing business owners they could take workers' tips.
Building trust in these tools will be paramount, which includes holding vendors and city officials accountable for mishaps like these.
Smart transportation: Multi-modal transportation networks, electric and autonomous vehicles, and the rise of intelligent infrastructure are all factors categorizing the ongoing shift toward smarter transportation networks, some of which can be supported by AI/ML.
China has been an early adopter of smart transportation technologies, which have been driven by the central government since as early as 2016. The private sector has also played a role. Hangzhou-based e-tail giant Alibaba has operated its City Brain offering since 2017, which aggregates data from traffic infrastructure, vehicles and mobile phones to optimize traffic flow and inform decision-making. The platform has been deployed in several cities across Southeast Asia.
AI will play a role in intelligent transportation systems and smart transportation networks through the continued support of horizontal technologies such as data fusion and computer vision (which can enable object detection and license plate recognition), as well as through applications like fleet management, optimized public transportation routing, and intelligent intersections.
City modeling and digital twins: Cities and other public sector organizations turn to digital twins to create an interactive copy of a city or site for simulations, planning and decision-making. A digital twin is a high-fidelity virtual representation that can integrate geographic information system (GIS) data, data from IoT devices, computer-aided design (CAD) files, existing records and more.
Industrial vendors are increasingly working with public sector organizations to create digital twins of critical infrastructure, including power plants, water and sewer systems, transportation networks and ultimately an entire city. With a digital twin, a city can understand which neighborhoods may flood during storms, and better prepare a response plan ahead of time. Other uses include mapping power grids or fiber networks to respond to outages faster and identify impacts.
Beyond extending their scope to encompass a city more holistically, city digital twins can become the basis of two-way interactions that facilitate a learning feedback loop. While IoT sensors collect data to offer insight into predictive analytics, such as when a heating, ventilation and air conditioning (HVAC) system in a city building is likely to fail, adapting this data into an interactive, intelligent assistant could enhance user experience and insights.
What may have once been delivered as an alert could instead contextualize broader data and deliver it in an interactive format. This interaction will move both ways — from a chatbot/intelligent assistant informing stakeholders of alerts, and from stakeholders making actionable requests that can be operationalized.
Combining IoT data collection with AI analytics and insights could usher in a digital transformation for cities of all sizes and geographies. But while these technologies have the potential to solve some of the public sector's most pressing issues around sustainability and safety, they are not without their risks. Data transparency, community engagement and digital equity should be paramount when cities and municipalities deploy smart city initiatives and work with vendors in the space.
For smart cities to achieve their potential, they will need both financial and popular support. That will depend on them proving their worth, and allaying fears of potential detriments. Doing so will require careful stewardship by government and private sector actors, and transparent discussion of the pros and cons of AI-powered enhancements.
Content Type
Research Council Theme
Director, Analytical Innovation