AI urbanism: opportunities and challenges

The world is becoming steadily more urbanized. Today, 56% of the world's population lives in cities, and by 2050 it’s forecast that seven out of ten people will live in cities.

As more people move to cities, the challenges associated with urban development and planning are ever more pressing for governments and urban development specialists. Traditional urban planning relied on historical data and the spread of demographics to plan and manage cities. However, as the population increased steadily, it became more difficult to predict future needs along with their complexities.

The emergence of Artificial intelligence (AI) technology has significantly facilitated urban planning and development. For instance, by leveraging Machine Learning (ML) technology, urban experts can analyze large volumes of historical datasets to predict future trends in urban development and identify possible challenges.

In this article, we’ll look at the challenges of using AI in urban planning. Neom, a future smart city in Saudi Arabia with a size almost 33 times the size of New York and an estimated cost of $500 billion, makes for a great example.

However, before we start, let’s check out some of the benefits of using AI in urban development and planning.

The benefits of AI Urbanism

There are numerous benefits to leveraging AI in urban planning and smart city management. For instance, many problems associated with current city planning can be mitigated by using AI to make clever infrastructure.

Smart infrastructure

AI can help in different areas of city management, such as:

• Smart waste management: Putting internet of things (IoT) sensors in street trash bins to automatically check when they’re full.
• Smart traffic control systems: AI algorithms can analyze real-time traffic data from sensors, GPS devices, and traffic cameras to predict traffic congestion and optimize traffic flow accordingly.
• Smart air quality: Measure air quality by using thousands of sensors spread across the city. In Barcelona, scientists have invented an AI model that uses machine learning to discover the urban areas generating the highest amount of nitrogen dioxide (NO2).
• Street Lighting: AI street lighting systems can adjust their brightness according to many criteria, such as inactivity periods and the light of the surrounding environment.

Urban planning

Machine learning techniques can be applied to solve different challenges in urban planning. The most significant applications involve leveraging machine learning algorithms to analyze satellite images to identify and map different land-use and land cover types within urban areas. This process involves using sophisticated algorithms that automatically classify and interpret satellite imagery to blueprint features such as vegetation, water bodies, and built-up areas.

Better public services

Powered AI chatbots can provide 24/7 services for citizens and make their lives more comfortable. For example, people can request information about public transport and traffic congestion, book appointments with city officials, or get information about some public services' opening hours.

Predictive capabilities

ML technologies can simulate the future expansion of cities before it happens and forecast the needed infrastructure (roads, electricity, green space, car parks, and other public utilities) over the years.

Cybersecurity challenges faced by smart cities

Smart cities emerged as a natural development of traditional urban development. They leverage different digital data technologies, AI, and connectivity to surpass the numerous problems faced in conventional cities.

Various countries aim to leverage AI to plan and manage their future smart cities. Nevertheless, the increased reliance on AI technology to power future cities will raise serious cybersecurity challenges. For instance, cyberattackers can execute various attacks to conduct malicious actions in the planning phase of the smart city and later attack its digital infrastructure and connectivity.

We can look at the Saudi smart city, Noam, in the same context. Saudi aims to build a modern city that leverages the highest technological advancement (some of these technologies are still under development in Western countries!) to create a sustainable environment powered by the latest AI and ML technologies. However, while being a pioneer in adopting the latest Hi-Tech has many competitive advantages, it will also make your project susceptible to numerous cybersecurity threats.

What cyber threats will Noam City face?

People will heavily depend on digital solutions in smart environments to perform their daily duties. They will frequently need to provide and share their sensitive data, such as location, health, and financial information, with online service providers and other entities providing services to them.

They also need to use digital communication mediums to exchange data with digital services. Automating many tasks in smart cities will allow service providers, city governments, and third-party vendors to collect vast amounts of sensitive data about people. For example, the intelligent waste bins will collect information on when and how frequently you throw your garbage. Sensors in the streets will record your transportation patterns and where you have been all day, and wearable medical sensors will record all your health indicators and send them to the health provider to monitor your health in real time. These were a few examples of living in a smart city.

What makes a cyberattack in a smart city environment so serious is that everything is connected in some way, and it is too difficult to isolate and contain security incidents. The interconnected world in smart cities will cause any sudden outage in communication technology or in a core service to halt other services, which can have negative consequences in different areas.

Let’s look at the main cyber threats that Noam – and similar smart cities – may be subject to:

Data bias risk

ML models used to power smart city development and planning are trained on massive datasets acquired from various sources, such as historical records, geospatial systems, and IoT sensors.

There are different cyberattacks against ML models. The poison attack is the most prominent one in the context of smart city planning. Threat actors may try to poison training data with biased information in the training phase to make the model give manipulated responses based on the threat actors' malicious objectives.

For example, a poisoned ML model may suggest improper resource allocation or unsafe infrastructure recommendations (e.g., putting public infrastructure areas in the wrong places or spreading them in a way that can not serve all communities equally).

Risk of cyberattacks

Smart cities depend mainly on digital infrastructures to provide key public utilities to the population. For instance, power grids, water treatment facilities, transportation networks, and communication systems will all depend on AI and automated systems in a smart city environment.

A cyber attack against any of these services or communication technologies will have catastrophic consequences (e.g., traffic chaos, poisoning water supplies, blackouts) on the entire city's operations.

The most prominent cyberattacks against smart cities will be ransomware and denial of Service (DoS) attacks.

Privacy issues

In smart cities, there will be an extensive collection of inhabitants' sensitive data from various parties, including government agencies, service providers, and numerous AI applications powering daily services.

However, the governance framework for managing this data remains vague. For instance, who has the access rights to personal data collected from millions of IoT sensors distributed across the city? Other concerns arise regarding the storage, processing, and access control mechanisms to protect and govern accessing this data.

For example, who will be authorized to access the vast amounts of personal data collected by IoT sensors in smart cities? Will this be limited to government agencies, service providers, or other entities? How will this data be securely stored, processed, and consumed to ensure privacy and data protection? These questions underscore the critical importance of addressing privacy issues in developing and deploying smart city technologies.

While no technological solution can guarantee complete security, the widespread adoption of AI technologies in urban planning and management offers significant advantages. However, as with all technology advancements, it also introduces security risks that must be addressed adequately to leverage AI securely in this context.